Skip to main content
Internet Archive's 25th Anniversary Logo

Full text of "Class and class conflict in industrial society"

See other formats

and Class Conflict 
in Industrial Society 

Ralf Dahrendorf 

^One of the most important efforts to 
reorient the approach of modern sociology 
. . . A major contribution to social theory." 

— Seymour Martin Lipset 


Stanford University Press 


Class and Class Conflict 
in Industrial Society 

PROF. iMni 



and Class Conflict 

in Industrial Society 



This work originally appeared in Ger- 
many in 1957 under the title Soziale 
Klassen und Klassenkon fltkt in der in- 
dustriellen Gesellschaft and has been 
translated, revised, and expanded by the 

Stanford University Press 

Stanford, California 

© 1959 by the Board of Trustees of the 

Leland Stanford Junior University 

Printed in the United States of America 

Cloth SBN 8047-0560-7 

Paper SBN 8047-0561-5 

First published 1959 

Last figure below indicates year of this printing: 

78 77 76 75 74 73 72 

(top". Y" 


and our common friends of the 

"Thursday Evening Seminar** 

at the London School of Economics (igß2—ig^/f) 

and to 

my fellow fellows 

at the Center for Advanced Study 

in the Behavioral Sciences ( I9S7~^95^) 

Preface to the First (German) Edition 

In the middle of the twentieth century, the sociologist finds him- 
self in an awkward position. While he is just beginning to lay and 
secure the foundations of his discipline, an impatient public demands 
with increasing urgency both immediately applicable and comprehen- 
sive solutions from him. Following almost every sociological confer- 
ence, confident journalists charge sociologists with being either igno- 
rant of practical problems or incapable of solving them. Like an angry 
creditor, the public pursues the sociologist's every move in order to 
lay its hands on every penny he may produce. Is it surprising that 
under these conditions many a sociologist has begun to forge currency.? 
The public deserves no better j but unfortunately the false currency, 
the overly rapid as well as the all-embracing solutions harm the prog- 
ress of sociology as well. They remove sociological discussion from 
the sphere of scholarly criticism which inspires ever new efforts to 
that of a sterile quarrel of opinions. It is therefore necessary to make 
a decision. 

The sociologist certainly is a debtor of the society in which he lives 
in a way unparalleled in most other disciplines of scholarship. But 
this debt merely forces him to choose the subjects of his research in 
such a way that its results — if any — might contribute to informing 
society about itself. This is all. In developing his theories, methods, 
and techniques, the sociologist is bound exclusively by the rules of 
all scholarship, which demand from him accuracy and adherence to 
that pedestrian path of science which nobody else can take for him 
and which no magic force can abridge. Inevitably, this path involves 
byways and detours j it is a long way, and its destination may disap- 
point his and others' expectations j perhaps somebody else soon shows 
that the path chosen was altogether wrong. But if an ill-advised 
public does not understand the process of scientific inquiry and de- 
mands more, the sociologist may and must be sufficiently proud and 
confident to defend his scholarly responsibility in face of a miscon- 
ceived obligation to society which is all too often informed by little 
more than a desire to please. Htc RhoduSy hie salta! 

viii Preface to the First (German) Edition 

These remarks are in place at the beginning of a study, the sub- 
ject of which is as extensive as its results are tentative, modest, and 
in need of supplementation. The attempt is made here to tackle a 
problem, which has for a long time been strangely neglected, with 
partly new and partly more refined means. To many it may sound 
surprising if I call a problem neglected about which the present 
study lists more than two hundred bibliographical references which 
could easily be doubled and trebled. Probably the word "class" 
belongs to the most frequently used words of sociology. But I am not 
concerned here with the word. I should not hesitate to replace it by 
a better one if I could find such} moreover, it will appear less fre- 
quently in the present investigation than might be expected. I am 
concerned with a problem, namely, with the puzzling fact that social 
structures as distinct from most other structures are capable of pro- 
ducing within themselves the elements of their supersession and 
change. Social structures not only are subject to change but create 
permanently and systematically some of the determinant forces of 
their change within themselves. Among these forces certain groups 
are paramount, the conflict of which may lead to modifications of 
existing values and institutions. I shall attempt to show in the present 
study how these groups and the processes to which they contribute 
can be identified theoretically and analyzed empirically. 

Perhaps a word of explanation is necessary as to why I have given 
a study of this problem the title Class and Class Conflict. At least 
one great sociologist, Karl Marx, has used the concept of class in 
the context intimated by the foregoing remarks. It is undeniable 
that not very many have followed Marx in this. Little more than 
a dozen (if important) sociologists who understand Marx's and our 
problem as one of class will be mentioned in the course of this study. 
Moreover, we shall have to subject the approaches of Marx and 
most later sociologists to severe criticism which often leads to the 
conclusion that they are vague, imprecise, incomplete, or even unten- 
able and erroneous. The overwhelming majority of sociologists since 
Sombart and Max Weber have associated the concept of class with 
other types of problems, especially with those of social stratification. 
A regrettable chain of circumstances seems to have committed both 
the original meaning of the concept of class and the problem of its 
first use to oblivion. All these factors can hardly serve to justify the 
attempts to revive both the problem and the concept of class in their 
original definition. However, so far as the problem is concerned, no 
justification is necessary, and with respect to the concept I shall try 

Preface to the First (German) Edition ix 

to point out that the situation is not quite as hopeless as it may seem 
at first sight. To anticipate but one argument here: There is, in soci- 
ological terminology, a useful alternative for the misunderstood 
concept of class, i.e., the term "stratum," whereas for the well- 
understood concept of class a substitute has not yet been found. 

For two reasons one can predict with some confidence that the 
present study will be misunderstood. One of these rests with the 
strict distinction of "class" and "stratum" and their respective heu- 
ristic purposes. By stratum I shall understand a category of persons 
who occupy a similar position on a hierarchical scale of certain situa- 
tional characteristics such as income, prestige, style of life. "Stratum" 
is a descriptive category. By contrast, the concept of class is an ana- 
lytical category which has meaning only in the context of a theory 
of class. "Classes" are interest groupings emerging from certain 
structural conditions which operate as such and effect structure 
changes. The confusion of these two concepts and spheres of analy- 
sis is so complete that I cannot hope to eliminate it entirely by this 
first attempt at clarification, even if I should have succeeded in sepa- 
rating class and stratum convincingly and consistently. I must accept 
the misunderstanding which is possible, even probable here, just like 
another one which goes even deeper and touches upon the patheti- 
cally preliminary discussion of the possibility of a sociological science. 
I ask the reader's indulgence if I refrain here from a general con- 
sideration of this subject and instead refer to the present study 
itself as a testimony to my conception of sociology. There is but one 
aspect of the problem which I should like to mention in advance, 
even as it is going to increase rather than mitigate misunderstand- 
ings: If in this study I speak of "theory," "hypothesis," "empirical 
test," "refutation," and "science," I use these terms in the strict 
sense of the methodological characteristics of an empirical discipline. 
At least logically, physics, physiology, and sociology are subject to 
the same laws — whatever may render one or the other of these dis- 
ciplines empirically preferable in terms of exactness. I cannot see 
why it should not be at least desirable to try to free sociology of 
the double fetters of an idiographic historical and a meta-empirical 
philosophical orientation and weld it into an exact social science with 
precisely — ideally, of course, mathematically — formulated postu- 
lates, theoretical models, and testable laws. The attempt must be 
made; and although the present study remains far removed from 
its satisfactory completion, I want it to be understood in terms of 
such an attempt. 

X Preface to the First (German) Edition 

Generalizing theoretical formulation and its empirical test are 
balanced in the present investigation. With R. K. Merton I regard 
"theories of the middle range" as the immediate task of sociological 
research: generalizations that are inspired by or oriented towards 
concrete observations. However, the exposition of the theory of 
social classes and class conflict stands in the center of this investiga- 
tion. The resume of Marx's theory of class, the largely descriptive 
account of some historical changes of the past century, and the criti- 
cal examination of some earlier theories of class, including that of 
Marx, lead up to the central theoretical chapters j with the analysis 
of post-capitalist society in terms of class theory a first empirical test 
of my theoretical position is intended. The whole investigation re- 
mains in the "middle range" also in that it is, as its title indicates, 
confined to industrial society. 

Many suggestions and stimulations which have gone into the 
present study originated in discussions in a small informal group of 
younger sociologists from diverse countries at the London School of 
Economics in the years 1952-54. This group, which called itself the 
"Thursday Evening Seminar," although it often continued its dis- 
cussions until Friday morning and met on other days as well, not 
only occupied itself with many of the specific questions of this study 
— such as Marx, Parsons, the whole problem of interest groups — 
but displayed a conception of sociology and its task which I hope 
to have upheld throughout this study. Within the "Thursday Eve- 
ning Seminar" and since, the stimulation of numerous conversations 
with Dr. D. Lockwood, Lecturer in Sociology at the London School 
of Economics, has, above all, furthered the progress of my own in- 
vestigation into class theory. In the hope that the provisional result 
of these investigations may provide a useful basis for critical dis- 
cussion I dedicate this study to David Lockwood and with him to 
our common friends of the London years. 

Scheint (Saar) 
S-pring igs7 

Preface to the Revised (English) Edition 

In every sense but one, this study is an essay even in its revised 
version. It is tentative, incomplete, open to criticism at many points, 
and, I hope, stimulating^ but it is also longer than the rules of 
essay writing would permit. Despite its length, I wish to empha- 
size the exploratory nature of my attempt to tackle problems of 
social conflict concerning total societies. By and large, recent devel- 
opments of sociology have been characterized by two related features. 
Firstly, there has been a strong concern for the conditions of "equi- 
librium" in "social systems." Stimulated by anthropological research, 
an image of society has gained prevalence in sociological thinking 
which emphasizes the elements of functional coordination, integra- 
tion, and consensus in units of social organization. The attempt to 
evolve testable theories and applicable conclusions has led, secondly, 
to an ever-growing interest in comparatively small "social systems" 
such as communities, enterprises, and, above all, small groups. Both 
these concerns of contemporary sociological analysis are, to be sure, 
important, and have proved fruitful. At the same time, however, 
they have led many to abandon completely such other subjects of 
sociological analysis as did not seem to fit in with the general trend. 
As a result, there is today a considerable need for reorienting socio- 
logical analysis to problems of change, conflict, and coercion in social 
structures, and especially in those of total societies. The interest in 
total societies, as well as in their historical dimension, is of course 
as old as sociology itself. Yet their neglect in recent decades makes 
a study like this one a venture into unmapped areas of inquiry — a 
venture which is guided not so much by the hope of comprehensive 
and final results as by the intention of challenging others to follow, 
criticize, and explore other avenues of discovery. 

From the reviews of the original German version of this study 
I have learned with some pride that it has in fact achieved at least 
one of its ends: it has stimulated critical discussion. It seems to me 
that few things are more deadly to the progress of knowledge than 

xii Preface to the Revised (English) Edition 

the deterioration of book reviews into advertisements. Lack of con- 
troversy means lack of interest, of stimulation and advance. I deem 
myself lucky to have escaped this pathetic fate. Most of the criti- 
cism of my study has been concerned with the theory of conflict 
whose rudiments are presented in the second half of this book. It 
has converged on five admittedly problematic points: (i) the reten- 
tion of the term "class" for a theory that dispenses with the histori- 
cal reference to antagonisms between "bourgeoisie" and "prole- 
tariat" 5 (2) the "definition" of the crucial concepts of power and 
authority; (3) the separation and, in part, the confrontation of in- 
dustrial and political class conflict j (4) the application of conflict 
theory to the analysis of contemporary society, especially to its "rul- 
ing classes" j (5) the absence of a grounding of this analysis on "large- 
scale inquiries of a more ^practical' kind." There have also been mis- 
understandings and more dogmatic (often Marxian) criticisms of my 
study J but the five points mentioned indicate comments and objec- 
tions which seem to me of particular importance. I do not propose to 
argue with my critics at this point. Instead, I present the revised 
edition of this essay, in which I have incorporated many suggestions 
offered by reviewers and other readers, have explicitly rejected others, 
and have tried to clarify parts that tended to cause misinterpretations. 
Where the revised version has improved on the original, this is due 
to no small extent to the critical comments of colleagues reviewing 
my book. 

The revised edition differs from the original in many respects j 
in fact, the author feels — with what probably is but a sign of lack 
of detachment from his own work — that it is a completely new 
book. For the sake of clarity of purpose, I have now divided the 
whole into two parts. Part One is concerned with a critical exami- 
nation of facts and theories relating to the problem of class. It deals 
above all with the doctrine of Marx, its empirical refutation and 
theoretical supersession. All chapters of this part have been ex- 
panded and partly rewritten j one of them — Chapter II — shares 
with its equivalent in the original little more than the title; other- 
wise hardly a word has remained the same. Part Two presents my 
attempt to approach conflict analysis both in terms of abstract con- 
siderations (Chapters V and VI) and with reference to post-capi- 
talist society (Chapters VII and VIII). Here, two entirely new 
chapters have been added, incorporating refinements of the theory 
as well as extensions of its application, while the remaining two have 
undergone substantial revision. Careful scrutiny of the emphases in 

Preface to the Revised (English) Edition xiii 

the work of revision would reveal that in the years since this book 
was originally written my interest has shifted from problems of 
industry to those of politics. Most of the additions and changes in 
the present edition have been stimulated by the desire to render con- 
flict theory applicable to the analysis of the political process both 
in totalitarian and in free societies of the present. A number of new 
books which either appeared or came to my knowledge since the date 
of the first publication of this study have been incorporated in the 
substance of the investigation as well as in its Bibliography. 

Finally, an important difference between the original and the 
revised version of this study consists in the fact that the former was 
written in German, the latter in English. I should like to empha- 
size the expression "written in English," for this is not, strictly speak- 
ing, a translation. The author translating his own book has the su- 
preme advantage of being free with respect to his text. He can 
reformulate, change, even leave out at will a phrase that sounds 
reasonable in one but awkward in the other language. This may be 
worrying to philologists, but it is gratifying both to the author and 
— I hope — to his readers. Also, the author translator is not held 
up by problems of interpretation j presumably he knows what he 
meant by his statements. (In fact, I found the task of "transla- 
tion" an acid test for ambiguities of thought and formulation, and 
therefore a welcome opportunity for rendering many a passage 
more precise.) I fear, however, that the advantages of "transla- 
tion" by the author are in this case more than balanced by the 
considerable disadvantages accruing from the fact that I have "trans- 
lated" this book into a language other than my mother tongue. If 
the resulting text is at all readable, this is due in no small extent 
to the careful and competent editing on the part of the publisher, 
Mr. Leon Seltzer, and his editors. It is above all due to the assist- 
ance of my wife who shares this book v/ith me both in the tangible 
sense of having put many hours of work into correcting the lan- 
guage (her mother tongue being English), typing, offering sugges- 
tions, and listening, and in numerous less tangible yet even more 
important ways. 

I am profoundly grateful also to Mr. G. Fleischmann, my as- 
sistant and collaborator at the Akademie für Genieinw'irtschaft in 
Hamburg, for having put aside his own work in order to help me 
in preparing and editing the manuscript, compiling the Bibliography, 
checking quotations, preparing the Index, and discussing with me 
many substantial points and arguments raised in this study. 

xiv Preface to the Revised (English) Edition 

I had dedicated the original edition of this study to my friend 
Dr. David Lockwood and to our common friends at the London 
School of Economics, because it was with and through these that I 
received the impulses that made me undertake this investigation. 
Whatever new ideas and analyses there may be in this revised edi- 
tion are inspired largely by many friendly discussions with fellow 
fellows at the Center for Advanced Study in the Behavioral Sciences 
in 1957-58, both inside and outside the "Conflict Seminar." With 
especial gratitude I record here my debt to Joseph Ben-David, 
Ph.D., John Bowlby, M.D., Professor Frank Newman, LL.D., 
Professor Fritz Stern, Ph.D.j and I should like to add Professors 
R. Bendix and S. M. Lipset of the University of California at Berke- 
ley. As an outward acknowledgment of my indebtedness to these 
colleagues and friends as well as to many other co-fellows at the 
Center I have extended to them the dedication of this book. 


Spring igs9 


Preface to the First (German) Edition vii 

Preface to the Revised (English) Edition xi 


I. Karl Marx's Model of the Class Society 3 

The social etymology of the concept of class, 3 Consequences of 
industrialization, 4 Marx's theory of class, 8 The problem, 9 
Two false approaches, 10 Property and economic power, 11 
Relations of production, class situation, and political power, 12 
Class interests, 14 Class organization and class struggle, 16 The 
classless society, 1 8 Sociological elements of Marx's theory of class, 
1 8 Philosophical elements of Marx's theory of class, 27 Marx's 
image of the capitalist class society, 32 

II. Changes in the Structure of Industrial Societies Since Marx 36 

Capitalism versus industrial society, 36 Ownership and control, 
or the decomposition of capital, 41 Skill and stratification, or the 
decomposition of labor, 48 The "new middle class," 5 I Social 
mobility, 57 Equality in theory and practice, 61 The institu- 
tionalization of class conflict, 64 Capitalism cum industrial so- 
ciety, 67 

III. Some Recent Theories of Class Conflict in Modern Societies 72 

Refutation is not enough, 72 The dilution of the concept of class, 
74 On and off the party line, 77 Capitalism, socialism, and 
social classes, 84 The managerial and the clerical revolutions, 87 
Class society without class conflict, 93 Class society in the melting- 
pot, 97 Citizenship, equality, and social class, 100 The new 
society, 109 Unsolved problems, 114 

IV. A Sociological Critique of Marx 117 

Sociology and the work of Marx, 117 Social structure and social 
change: Marx sustained, 119 Social change and class conflict (i): 
Marx sustained, 124 Social change and class conflict (11): Marx 
rejected, 126 Class conflict and revolution: Marx rejected, 130 
Social classes and class conflict: Marx rejected, 133 Property and 
social class: Marx rejected, 136 Industry and society: Marx re- 
jected, 141 Social roles and their personnel: Marx supplemented, 
144 The concept and theory of class, 150 

cvi Contents 


V. Social Structure, Group Interests, and Conflict Groups 1 57 

Integration and values versus coercion and interests: the two faces of 
society, 157 Power and authority, 165 Latent and manifest 
interests, 173 Quasi-groups and interest groups (i): theoretical 
conditions of conflict group formation, 179 Quasi-groups and in- 
terest groups (n): empirical conditions of conflict group formation, 
182 A note on the psychology of conflict groups, 189 "Elites" 
and "ruling classes," 193 "Masses" and "suppressed classes," 198 
Classes or conflict groups?, 201 

VI. Conflict Groups, Group Conflicts, and Social Change 2o6 

The "functions" of social conflict, 206 Intensity and violence: 
the variability of class conflict, 210 Pluralism versus superimposi- 
tion: contexts and types of conflict, 213 Pluralism versus super- 
imposition: authority and the distribution of rewards and facilities, 
215 Mobility versus immobility: the "classless" society, 218 
The regulation of class conflict, 223 Group conflict and structure 
change, 231 The theory of social classes and class conflict, 236 

VII. Classes in Post-Capitalist Society (I): Industrial Conflict 241 

Capitalist society in the light of the theory of group conflict, 241 
Do we still have a class society?, 246 The authority structure of 
the industrial enterprise, 248 Industrial democracy, 257 The 
institutional isolation of industry and industrial conflict, 267 Em- 
pirical consequences of the theory of institutional isolation of indus- 
trial conflict, 272 Industrial conflict: trends and countertrends, 

VIII. Classes in Post-Capitalist Society (II): Political Conflict 280 

How people see society, 280 The authority structure of the po- 
litical state, 289 Bureaucratic roles and political authority, 295 
The ruling class, 301 Political democracy, 307 Totalitarian 
societies vs. free societies, 314 

Bibliography 3 1 9 

Subject Index 329 

Author Index 333 


The Marxian Doctrine in the Light of 
Historical Changes and Sociological Insights 


Karl Marx^s Model of the Class Society 


The concept of class has never remained a harmless concept for 
very long. Particularly when applied to human beings and their social 
conditions it has invariably displayed a peculiar explosiveness. The 
logician runs no risk in distinguishing "classes" of judgments or cate- 
gories j the biologist need not worry about "classifying" the organisms 
with which he is concerned — but if the sociologist uses the concept of 
class he not only must carefully explain in which of its many meanings 
he wants it to be understood, but also must expect objections that are 
dictated less by scientific insight than by political prejudice. As Lipset 
and Bendix have stated: "Discussions of diflFerent theories of class are 
often academic substitutes for a real conflict over political orienta- 
tions" (55, p. 150).' 

We shall have to show where this impermissible and unfortunate 
confusion of judgments of fact and value originates in this case, and 
we shall have to find ways and means to weld the concept and theory 
of class into useful tools of sociological analysis without evaluative 
overtones. However, for the time being we have to resign ourselves 
to the fact that using the concept of class may cause misunderstandings 
of many kinds. 

Evaluative shifts of meaning have accompanied the concept of 
class throughout its history. When the Roman censors introduced the 
word classis to divide the population into tax groups, they may not 
have anticipated the eventful future of this category. Yet even their 
classification implied at least the possibility of evaluative distinctions: 
on the one end of their classification were the assidui, who might well 
be proud of their 100,000 as; on the other end were the proletarii, 
whose only "property" consisted in their numerous offspring — proles 
— and who were outdone only by the lumpenproletariat of the capite 
censi, those counted by their heads. Just as the American term "in- 

^ Figures in parentheses refer to the corresponding numbers in the bibliography 
at the end of the volume. 

4 Marx^s Model 

come bracket," although originally no more than a statistical category, 
touches upon the most vulnerable point of social inequality, it was true 
for the classes of ancient Rome that they divided the population into 
more than statistical units. "The movie was classy," teen-agers say, 
meaning "high-class," "first-class." Similarly, to say that some Ro- 
man was classis or classicus meant that he belonged to the fnma classis, 
to the upper class — unless he was explicitly described as a "fifth-class" 
proletarian. Since Gellius we know the adjective classicus in its appli- 
cation to "first-class" artists and works of art, a usage which survives 
in our word "classical" and was eventually related to the authors of 
the term themselves and their times: they lived in "classical" an- 

When more recently sociologists remembered the word, they nat- 
urally gave it a slightly different connotation. Initially the word 
"class" was used — for example, by Ferguson (2) and Millar (15) 
in the eighteenth century — simply to distinguish social strata, as we 
should say today, by their rank or wealth. In this sense the word 
"class" can be found in all European languages in the late eighteenth 
century. In the nineteenth century the concept of class gradually took 
on a more definite coloring. Adam Smith had already spoken of the 
"poor" or "labouring class." In the works of Ricardo and Ure, Saint- 
Simon and Fourier, and of course in those of Engels and Marx the 
"class of capitalists" makes its appearance beside the "labouring class," 
the "rich" beside the "poor class," the "bourgeoisie" beside the "pro- 
letariat" (which has accompanied the concept of class from its Roman 
origins). Since this particular concept of social class was first applied 
in the middle of the nineteenth century, its history has been as event- 
ful as that of the society for which it was designed. However, before 
we embark on a critical journey through this history it appears useful 
to survey the meaning and significance of the "classical" concept of 
class as it was formulated by Karl Marx. 


"The history of the working class in England begins with the last 
half of the past [eighteenth] century, with the invention of the steam 
engine and the machines for manufacturing cotton," wrote the young 
Engels in 1845 (i, p. 31). With the industrial revolution also, the 
history of the concept of class as a tool of social analysis began. Earlier, 
the concepts of "class" and "rank" could be interchanged as by Fergu- 
son and Millar j indeed, that of "rank" could be preferred. The super- 
ficial observer at least was above all struck by "distinctions of rank" 

Marx's Model 5 

in late feudal society.^ In the emerging industrial society, however, 
rank and social position gave way to much cruder distinctions. As capi- 
tal, property became transformed from a symbol of rank to an instru- 
ment of power growing steadily in strength and effectiveness. Much 
as nobility and small independent peasants might resent it, both be- 
came witnesses and victims of the disappearance of an old and the 
emergence of a new social order, before which all well-tried categories 
of understanding and explanation failed. 

The history of the industrial revolution and its immediate conse- 
quences is too well known to be repeated here. However, one aspect 
of this history appears essential for our discussion. Wealth and pov- 
erty, domination and subjection, property and propertylessness, high 
and low prestige — all these were present before the industrial revolu- 
tion as afterward. Thus it might appear as if all the industrial revo- 
lution effected was to replace old social strata by new ones: landowners 
and nobility by capitalists, laborers and small peasants by proletarians. 
This presentation, however, not only is oversimplified but overlooks 
the revolutionary character of the changes which accompanied indus- 
trialization. The difference between the early stages of industrial 
society in Europe and its historical predecessor was not just due to a 
change in the personnel of social positions j it was due above all to the 
simultaneous abolition of the system of norms and values which guar- 
anteed and legitimized the order of preindustrial society. The "dis- 
tinctions of rank" in preindustrial societies of even the eighteenth 
century rested as much on a myth of tradition, an intricate system of 
age-old, often codified rights and duties, as on the comparatively 
crude gradations of property, power, and prestige. Preindustrial so- 
ciety, of course, had also had its beginnings. Its claim to the legiti- 
macy of the present was also a product of history or, perhaps, an ide- 
ology. Yet when it was hit by the revolution of industry, this society 
had an order endowed by the patina of centuries with a special claim 
to legitimacy and a special solidity. The power of the landlord was 
not based on his having money, land, or prestige, but on his being a 

^ Of course, Ferguson and Millar understood by "rank" by no means only what 
we call "prestige" today. In fact, Millar's formulation sounds surprisingly "modern": 
"According to the accidental differences of wealth possessed by individuals, a subordi- 
nation of ranks is gradually introduced, and different degrees of power and authority 
are assumed without opposition, by particular persons, or bestowed upon them by the 
personal voice of the society." The difference indicated above and caused by the 
industrial revolution is rather a difference of perspective, which may be expressed by 
the terms "estate" and "class." 

6 Marx's Model 

landlord as his fathers had been for time immemorial. The condi- 
tions of the master craftsman, his journeymen and apprentices, and 
even that of the laborer resembled that of the landlord in their legiti- 
mation by the authority of tradition. In this sense, preindustrial 
society was what contemporary sociologists like to call, with a some- 
what doubtful expression, a "relatively static social order" (cf. Cox 
40, p. 467). 

Precisely these features were eliminated by the industrial revolu- 
tion.^ Surprisingly soon it created — to begin with, in England — two 
rapidly growing new strata, those of entrepreneurs and workers. 
There was no "precedent" for either, even if in England the Poor 
Laws mixed the old and the new poor in the same way the Crown 
mixed the old and the new aristocracy. Both these strata, "bour- 
geoisie" and "proletariat," which had grown up together and were 
tied to each other, had no tradition of rank, no myth of legitimacy, 
no "prestige of descent" (to quote Max Weber). They were charac- 
terized solely by the crude indices of possession and nonpossession, 
of domination and subjection. Industrial capitalists and laborers had 
no "natural," no traditional, unity as strata. In order to gain it, they 
had to stabilize and create their own traditions. They were, so to 
speak, nouveaux riches and nouveaux fcrnvres, intruders in a system of 
inherited values and messengers of a new system. And for these strata, 
bare of all traditions and differentiated merely by external, almost 
material criteria, the concept of "class" was first used in modern social 
science. In the analysis of these strata this concept became a sociologi- 
cal category. It is significant that in conversational German the word 
"class" is even today confined to the two strata of entrepreneurs and 
workers. Neither the nobility nor the professions nor the older groups 
of craftsmen and peasants are called classes. They are "estates" — a 
concept which in the case of the "middle estate" {Mittelstand) has 
been retained even for the newer groups of white-collar workers and 
civil servants.* An estate, however, is something else than a stratum or 

' A schematic sketch like the one attempted here obviously ignores local differ- 
ences as well as the gradual character of the emergence of industrial societies. All 
social historians of industrial development — from Weber (189) and Sombart (28), 
Tawney (187), and the Hammonds (175) to Bendix (138) and Jantke (178) in 
recent years — emphasize the gradual breakdown of the traditions of agrarian society. 
Concentration on the imaginary point of an "industrial revolution" can be justified 
only by the analytical purpose of these introductory remarks. 

* The significance of this German usage is of course only partly open to generali- 
zation. While on the one hand it documents the thesis here advanced about the his- 
torical context of the concept of class, it testifies on the other hand to the continued 

Marx's Model 7 

class, not only in everyday language but for the sociologist as well. 
**Status position' can be based on class position of a definite or indefi- 
nite kind. But it is not determined by class position alone: possession 
of money or the position of entrepreneur are not in themselves status 
qualifications, although they can become suchj propertylessness is not 
in itself status disqualification, although it can become such'* (M. 
Weber 330, p. 180). "A number of persons forming a social-status 
stratum more or less clearly delimited from other strata in customary 
or statutory law constitutes a social estate" (Cox 40, p. 467). An 
estate, as against a class or — to anticipate the result of later conceptual 
discussions at this point — an open stratum, is characterized by the very 
attributes which the bourgeoisie and the proletariat of early industrial 
societies were lacking: the sacred tie of tradition and the undisputed 
belief in the historically founded legitimacy of position. 

Thus the concept of class in its modern formulation is, if not the 
result of a definite historical situation, oriented toward and linked 
with such a situation. How difficult it is to extract the concept from this 
situation and apply it to other groups than bourgeoisie and proletariat, 
other societies than the industrializing societies of Europe, is shown 
most clearly by the long and still unconcluded sociological discussion 
of the work of Marx. The attempt to generalize the concept of class 
would of course hardly be worth the effort, if it were merely a name 
for social units like bourgeoisie or proletariat. In fact it is more. Since 
Marx, "class," "stratum," "rank," and "position" are no longer inter- 
changeable names for identical groupings. For even if Marx received 
most of his inspiration and material from the situation of English so- 
ciety half a century after the industrial revolution, this society was in 
a sense no more than an example for him, an illustration that served 
to test the usefulness of a more general approach. Since Marx devel- 
oped, on the basis of the concept of class, at least the rudiments of a 
theory of class, the model of a class society, his approach stands at the 
beginning of the considerations of this study. 

importance of preindustrial strata in German society-. The use of the term "middle 
class" in English is certainly no accident. 

° In translating Weber's term Stand, most translators have used the word "status." 
This — though not false — is misleading in that it does not convey the double meaning 
of the German Stand as "status" and "estate." In the passage quoted here, Weber 
undoubtedly meant to describe status in an estate context (rather than, for example, 
prestige status). This is only one example of the exigencies of translations — and of 
their creativity. By the very fact of misleading they can create terms that acquire 
a life of their own. 

8 Marx's Model 


There have been many and violent disputes about the interpre- 
tation of the work of Marx, but no commentator has seriously doubted 
the central importance of the theory of class for this work. Indeed, 
the greatness and fatality of his work become apparent in Marx's 
theory of class. In this theory, the three roots of his thought are 
joined. Marx adopted the word from the early British political econo- 
mists; its application to "capitalists" and "proletarians" stems from 
the French "utopian" socialists; the conception of the class struggle 
is based on Hegel's dialectics. The theory of class provides the prob- 
lematic link between sociological analysis and philosophical specula- 
tion in the work of Marx. Both can be separated, and have to be sep- 
arated, but in this process the theory of class is cut in two; for it is as 
essential for Marx's philosophy of history as it is for his analysis of 
the dynamics of capitalist society. 

Marx regarded the theory of class as so important that he post- 
poned its systematic exposition time and again in favor of refinements 
by empirical analysis. As a result we know it only by its application 
to concrete problems and by the occasional generalizations that occur 
throughout Marx's works. This may not be the least cause of the 
many controversies about the real meaning of Marx's concept and 
theory of class. Only recently the accounts of Geiger (46) and of 
Bendix and Lipset (36) have concluded these discussions, at least in 
"Western" sociology. It is not my intention, in presenting my inter- 
pretation of Marx's approach, to relight the fires of controversy. The 
following discussion of the concept and theory of class in the work 
of Marx, while not materially deviating from either Geiger or Bendix 
and Lipset, is designed to supplement their works and add some sub- 
stance to an investigation that is indebted to Marx even in its most 
radical criticisms of his work. 

Marx postponed the systematic presentation of his theory of class 
until death took the pen from his hand. The irony has often been 
noted that the last (52nd) chapter of the last (third) volume of Capi- 
tal, which bears the title "The Classes," has remained unfinished. 
After little more than one page the text ends with the lapidary remark 
of its editor, Engels: "Here the manuscript breaks off." However, 
for the thorough reader of Marx this is no reason for despair. If he 
wants to, he can complete this chapter for Marx — not exactly as Marx 
would have written it, of course, and not entirely without interpreta- 
tion either, but in any case without substantially adding to what Marx 
said himself. In the following section I shall try to do just this. By 

Marx's Model 9 

systematically ordering a number of quotations and connecting them 
to a coherent text I shall attempt to provide a basis and point of ref- 
erence for critical discussion without anticipating — beyond the selec- 
tion and ordering of the quotations — any interpretation.® 

The unwritten 52nd chapter of Volume III of Marx's Cafital 


It is the ultimate fur-pose of this work to reveal the economic laws 
of develofm^ent of modern society ( 1 2, 1, pp. 7 f .) . We are therefore 
not concerned with merely describing, much less regretting, existing 
conditions, but want to lay bare their revolutionary aspect. We have 
shown that the capitalist mode of production has become too restricted 
for its own forces of production. The revolution is near. But this 
revolution is not the product of economic forces of production or rela- 
tions of production, but of the people and groups that represent these 
economic formations. Of all instruments of production the greatest 
force of production is the revolutionary class itself (6, p. 188). 

For almost forty years we have emphasized the class struggle as 
the primary motive force of history^ and especially the class struggle 
between bourgeoisie and proletariat as the great lever of modern social 
change ( 1 1, p. 102). With the moment in which civilization be gins ^ 
production begins to be based on the antagonism between accumulated 
and direct labor. Without conflict^ no progress: that is the law which 
civilization has followed to the present day. Until now the forces of 
production have developed by virtue of the dominance of class con- 
flict (6, p. 80). And it always holds that a change in the relation of 
classes is a historical change (5, II, p. 475). 

Thus we have to determine in general what constitutes a class and 
how class conflict emerges and expresses itself. In a general investi- 
gation of this kind it is always assumed that real coftditions correspond 
to their conception, or, which is the same things that real conditions 
are presented only in so far as they express their own general type 
(12, III, p, 167). We are therefore not concerned with describing 
any one society, but with discovering the general laws which determine 
the trend of social development. 

// we observe a given country from the point of view of political 
economy, we have to start with its population, its distribution into 

® All quotations from Marx in the following section are in italics. Everything 
else is my text. 

10 Marx's Model 

classes, town, country, sea, the different industries, exfort and import, 
annual production and consumption, commodity prices, etc. (7, 
p, 256). But this method presents difficulties. It leads us astray if in 
our abstractions we do not find the way to the real and concrete, the 
real premise. Population is an abstraction if I ignore, for example, the 
classes of which it consists. These classes are again an empty word, if 
I do not know the elements on which they are based, e.g., wage labor, 
capital, etc. (7, p. 256). Thus our first question concerns the elements 
on which classes are based j and since modern bourgeois society is in 
fact our main subject (7, p. 237), we use it for the time being as an 

The owners of mere labor power, the owners of capital, and the 
landowners, whose respective sources of income are wage, profit, and 
rent — thus wage laborers, capitalists, and landowners — constitute the 
three great classes of modern society based on a capitalist mode of 

In England, modern society has undoubtedly advanced furthest 
and most classically in its economic structure. Even there, however, 
this class structure is not displayed in a pure form. Intermediate and 
transitional stages obliterate the borderlines there as everywhere (al- 
though incomparably less in the country than in towns). However, 
this does not matter for our investigation. It has been demonstrated 
that it is the per'^nanent tendency and law of development of the capi- 
talist mode of production to separate the means of production increas- 
ingly from labor, and to concentrate the separate m-eans of production 
m^ore and tnore in large groups — in other words, to transform labor 
into wage labor, and the means of production into capital. A t the same 
tvme^ land ownership tends to be separated from capital and labor, and 
to be converted into the type of land ownership corresponding to the 
capitalist mode of production. 

The question to be answered next is: What constitutes a class? 
And this results directly from the answer to the other question: What 
makes wage laborers, capitalists, and landowners the constituent forces 
of the three great social classes? ( 12, III, p. 941.) 


At first it is the identity of revenues and sources of inco^ne. They 
are three large social groups, whose components, i.e., the people of 
whom they consist, earn their living by wage, profit, and rent, i.e., by 
utilizing their labor power, capital, and land ownership. However, 
from this point of view, say, doctors and civil servants would also 

Marx's Model 1 1 

constitute two classes^ for they belong to two different social groups 
whose members^ inco?nes flow from the same source. The same would 
hold for the infinite fragmentation of interests and positions which 
the division of labor produces mnong workers as among capitalists and 
landowners (the latter ^ for example ^ into vineyard owners y field own- 
ers y forest owner Sy m^ine owner Sy fishing ground owners) (12, III, 
pp. 941 f.). 

This approach does not therefore lead to a fruitful definition. 
The same holds for a second approach frequently adopted in ex- 
plaining class differences and conflicts. The vulgar m,ind commutes 
class differences into ^^differences in the size of purses^* and class 
conflict into ^^ trade disputes.*^ The size of the purse is a purely quan- 
titative differencey by virtue of which two individuals of the same 
class can be opposed quite arbitrarily. It is well known that medieval 
guilds quarreled with each other ^^ according to trade. ^^ But it is equal- 
ly well known that modern class differences are by no m^eans based 
on "trade.** Rathery the division of labor has created very different 
types of work within the same class (5, II, pp. 466 f.). 

In both cases the essential point is overlooked: property, income, 
and source of income are themselves a result of the class structure, 
i.e., of the structure of economic conditions. Income and property 
are criteria belonging to the realm of distribution and consumption. 
However, the use of products is determined by the social relations 
of the consumerSy and these social relations themselves rest on the 
conflict of classes (6, p. 81). And since distribution is itself a prod- 
uct of productiony the kind of participation in production determines 
the particular patterns of distributiony the way in which people par- 
ticipate in distribution (7, p. 250). 

There is no property anterior to the relations of domination and 
subjection which obtain in production and in the political state, and 
which are far more concrete relations (7, p. 258) . Therefore we have 
to look for the elements of classes in production and in the power 
relations determined by it. 


The essential condition that determines the mode of production 
of an epoch, and that therefore provides the constituent element of 
classes as well as the momentum of social change, is property. The 
property question, relative to the different stages of development 
of industry y has always been the life question of any given class (5, 
p. 459)- 

1 2 Marx's Model 

However, this statement is open to misinterpretation. For the 
offosition of fro-pertylessness and froferty as such is indijferenty 
and not expressed in an active relation to its inner structure ^ i.e.y as 
a contradiction, so long as it is not comprehended as the opposition 
between labor and capital (3, p. 176). 

Even in this specification property is still an abstraction, an empty 
concept. In every historical epoch property has developed differently 
and under different social conditions. To define bourgeois property 
means no less than to describe all the social conditions of bourgeois 
production. The attempt to define property as an independent re- 
lation, a special category, an abstract and eternal idea, can be noth- 
ing but an illusion of metaphysics or jurisprudence (6, p. 169). 

Only if we understand property in the particular context of 
bourgeois society, i.e., as private ownership of the means of produc- 
tion, as the control of a minority over the wealth of a whole nation, 
do we in fact grasp the core of the antagonism existing in production 
and creating class conflict. The power of society thus becomes the 
private power of a private person (12, I, p. 138). 

The essential condition of the existence and domination of the 
bourgeois class is the accumulation of wealth in the hands of private 
persons, the formation and augmentation of capital; the condition 
of capital is wage labor (14, p. 89). Thus the existence of capital 
as well as wage labor, of the bourgeoisie as well as the proletariat, 
can be explained in terms of the one condition of the particular form 
of property in bourgeois society, i.e., ownership of the means of 

The authority relations within production which are given by 
the presence or absence of effective property, of control over the 
means of production, are of course not the class relations themselves. 
In order to determine these, we have to look for the consequences 
flowing from the relations of production and for the social antago- 
nisms based on these consequences. 


One important consequence of the relations of production has 
already been mentioned. The division of wealth in the sphere of 
distribution corresponds to the division of property in production. 
Thus a person's material condition of existence, or class situation, is 
based on his position in production. Economic conditions have first 
converted the mass of the population into workers. The rule of 

Marx^s Model 13 

cafital has created for this mass a common situation ( 6, p. 187). And 
in a way one can state: In so far as m-illions of families live under 
economic conditions which separate their way of life, their interestSy 
and their education from those of other classes and of'pose them to 
these y they constitute a class (8, p. 104). 

However, these economic conditions of existence are not in them- 
selves sufficient for the formation of classes. They are as such pas- 
sive, and although they produce the gaf between the life situations 
of worker and capitalist (12, I, p. 548), they do not produce a real 
antagonism. For in so far as there is between people in a common 
material condition, or life situation, a merely external contact — in 
so far as the identity of their interests does not 'produce a community y 
national associationy and political organization — they do not consti- 
tute a class. Such groups in a common situation are therefore unable 
to m,ake their class interest heard in their own name through a par- 
liam^ent or an assembly (8, p. 104). We shall have to return to this 

A second and infinitely more important consequence of the dis- 
tribution of property in production is that it determines the distri- 
bution of political power in society. Modern relations of production 
include the economic power of the owners of private property, the 
capitalists. And the political power of the bourgeois class arises from 
these modern relations of production (5, p. 455). Indeed it can be 
said that the modern state is but an association that administrates the 
common business of the whole bourgeois class (14, p. 83). 

In this sense, authority relations in production determine the 
authority relations of society in general. The specific economic form, 
in which unpaid surplus labor is pumped out of the immediate pro- 
ducers determines the relation of domination and subjection as it 
grows directly out of and in turn determines production. On this is 
based the whole structure of the economic com^nunity as it cofnes 
forth from the relations of productiony and thereby at the same time 
its political structure. It is always the immediate relation of the 
owners of the conditions of production to the imfnediate producers — 
a relation whose specific pattern of course always corresponds to a 
certain stage in the developfnent of labor and its social force of pro- 
duction — in which we find the final secrety the hidden basis of the 
whole construction of society y including the political patterns of sov- 
ereignty and dependencCy in shorty of a given specific form of gov- 
ernment { 12, III, p. 841). 

Finally, a third and parallel consequence of the distribution of 

14 Marx^s Model 

property in production is that it also shapes the ideas that mold the 
character of a period. On the different forms of froferty and the 
social conditions of existence a whole superstructure of various and 
peculiarly formed sentiments^ illusions^ modes of thought^ and con- 
ceptions of life is built. The whole class creates and forms these 
out of its Tnaterial foundations and the corresponding social relations 

(8, p. 37). 

We can say, therefore, that the ruling ideas of a period have 

always been nothing but the ideas of the ruling class (14, p. 93). 

In each efoch, the thoughts of the ruling class are the ruling thoughts ; 

i.e.y the class that is the ruling material 'power of society is at the same 

time its ruling intellectual power. The class that has the m-eans of 

m^aterial production in its control ^ controls at the same time the means 

of intellectual production (13, II, p. 37). 


We have seen that relations of property and authority constitute 
the basis of the formation of social classes. But we have not yet 
investigated the force that effects this formation. Classes do not 
exist in isolation, independent of other classes to which they are 
opposed. Individuals form a class only in so far as they are engaged 
in a comm^on struggle with another class (13, II, p. 59) j and the 
force that effects class formation is class interest. In a sense, class 
interests precede the formation of classes. Thus the German bour- 
geoisie stands in opposition to the proletariat even before it has or- 
ganized itself as a class in the political sphere (5, p. 469). The pro- 
letariat has, in the beginning of its development, certain common 
interests^ but it is nevertheless still an unorganized mass. Thus this 
mass is already a class in opposition to capital^ but not yet a class for 
itself (6, p. 187). 

By postulating class interests as preceding the classes themselves, 
we make it quite clear that class interests are not merely the random 
personal interests of one person or even many people. We are not con- 
cerned with what this or that proletarian or even the whole prole- 
tariat visualizes as a goal for the time being. Its goal and its histori- 
cal action are obviously and irrevocably predetermined by its own 
life situation as by the whole organization of contemporary bourgeois 
society (4, p. 207). Thus the shared interest of a class exists not only 
in the imagination^ as a generality y but above all in reality as the 
m,utual dependence of the individuals among whom labor is divided 
(13, II, p. 23). As in private life we distinguish between what a 

Marx's Model 1 5 

man thinks and says of himself and what he really is and doeSy so 
in historical struggles we must distinguish even more carefully the 
catchwords and fantasies of parties from their real organism and their 
real interests y their conception from their reality (8, p. 38). 

Class interests as "objective" interests subsuming the members 
of a class under a general force not only can differ from individual, 
personal interests, but can conflict with these interests. Although, 
for example, all members of the m^odern bourgeoisie have the same 
interest inasmuch as they form a class vis-a-vis another class, they have 
nevertheless o-p-posite^ contradictory interests as soon as they are con- 
fronted with each other (6, p. 140). This conflict of interests is not 
merely a possibility j it arises with a degree of necessity from the eco- 
nomic conditions of their bourgeois life (6, p. 140). For example, 
the conflict between the interest of the individual capitalist and the 
class of capitalists makes itself felt if the problem at hand is not the 
distribution of profits but that of losses, just as before the identity 
of interests found its practical realization through competition (12, 
III, p. 282). 

The substance of class interests, in so far as they are based on the 
economic positions of given groups, can be expressed in various ways. 
To begin with, the immediate interest of the proletariat is the wage, 
that of the bourgeoisie the profit j and here once again we have to 
distinguish the two great categories into which the interest of the 
bourgeoisie is divided — land ownership and capital (8, p. 38). From 
these immediate concerns, confined to the sphere of production, all 
further interests can be derived. As a society develops to its maturity, 
the originally divided interests become increasingly united. More 
and more it is a specific type of production, and of relations of pro- 
duction, which determines rank and influence of all other activities 
(7, p. 264). This means that two particular interests are increasingly 
articulated: the conservative interest of the ruling class, and the revo- 
lutionary interest of the oppressed class. Of all the classes with which 
the bourgeoisie is today confronted, only the proletariat is a truly 
revolutiormry class (14, p. 88). And a class in which the revolution- 
ary interests of society are concentrated, as soon as it has risen up, 
finds directly in its own situation the content and the material of its 
revolutionary activity: foes to be laid low; measures, dictated by 
the needs of the struggle, to be taken — the consequences of its own 
deeds drive it on. It makes no theoretical inquiries into its task 
(9, p. 42). 

On the basis of these class interests, in fighting to realize them 

1 6 Marx's Model 

or defend them, the groups determined by the distribution of prop- 
erty in production, and by the distribution of political power flowing 
from it, organize themselves into classes. 


The organization of classes follows the progress of conflicts 
within the sphere of production itself. Increasingly the collisions 
between the individual worker and the individual bourgeois assume 
the character of collisions between two classes. The workers start 
forming coalitions against the bourgeois; they join in order to m^ain- 
tain their wage ( 14, p. 87) . But the wage is, as we have seen, merely 
an undeveloped, prerevolutionary interest of the proletariat. This 
stage of class organization corresponds to a relatively early phase of 
capitalist development. As long as the rule of the bourgeois class 
had not organised itself fully ^ and had not acquired its fure political 
expression, the opposition of the other classes could not come forth 
in its pure form either , and where it did come forth, it could not take 
that dangerous turn which converts every struggle against govern- 
m^ent into a struggle against capital (8, p. 54). The development of 
the forces of production has to be far advanced for the formation of 
classes to be possible, because the Organization of the revolutionary 
elements as a class presupposes the complete existence of all forces 
of production which could possibly develop in the womb of the old 
society (6, p. 188). 

The formation of classes always means the organization of com- 
mon interests in the sphere of politics. This point needs to be em- 
phasized. Classes are political groups united by a common interest. 
The struggle between two classes is a political struggle (6, p. 187). 
We therefore speak of classes only in the realm of political conflict. 
Thus every movement in which the working class as such opposes 
the ruling class and seeks to destroy its power by pressure from with- 
out is a political move^nent. The atternpt, for example, to extort a 
limitation of working time in a single factory or trade, and from 
individual capitalists, by strikes, etc., is a purely economic fnovement ; 
but the movement to enforce legislation stipulating an eight-hour 
day, etc., is a political movement. And in this manner a political 
movement grows everywhere out of the isolated economic move- 
ments of the workers; i.e., it is a movejnent of the class in order to 
realize its interests in a general form, in a form that possesses uni- 
versal social constraining force (10, p. 90). 

Parallel with the political organization of classes there grows up 

Marx's Model 17 

a theoretical class-consciousness (12, I, p. 13), i.e., an awareness on 
the individual's part of the interests of his class generally. The posi- 
tive goals of the proletariat become evident and can be formulated 
by its theoreticians. As long as the 'proletariat has not sufficiently 
developed to organize itself as a class, as long as therefore the strug- 
gle of the proletariat with the bourgeoisie as yet has no political 
character J these theoreticians are merely Utopians who invent sys- 
tems in order to satisfy the needs of the oppressed classes (6, p. 142). 

Thus classes are political forces based on the relations of prop- 
erty and power. But although in principle every individual can be 
identified as a member of one of the above-named classes according 
to his share in property and power, it is quite possible that a man's 
actions will not always be determined by the class to which he belongs; 
hut these individual cases are as irrelevant to the class struggle as the 
defection of some noblemen to the Third Estate was to the French 
revolution (5, p. 467). 

This circulation among the classes or exchange between thetn ( 7, 
p. 266) is particularly evident in two stages of the organization of 
interest groups into classes. We find it in the first place, for example, 
in the United States of America, where although classes exist they 
have not yet become stabilized, but instead exchange and transfer their 
elements in continuous flux (8, p. 18). That is to say, we find this 
exchange in an early stage of class formation when the ruling class 
is still concerned with consolidating its power. And the fnore capable 
a ruling class is of absorbinz the best men of the olypressed class, the 
more solid and dangerous is its rule (12, III, p. 649). The second 
stage in which a certain exchange between the classes takes place is 
that immediately preceding a revolution. In times in which the class 
struggle approaches its decision, the process of disintegration within 
the ruling class and within the whole old society assumes such a vio- 
lent and glaring character that a small part of the ruling class re- 
nounces it and joins the revolutionary class, the class that carries the 
future in its hands. Just as earlier a part of the nobility went over 
to the bourgeoisie, now a part of the bourgeoisie goes over to the pro- 
letariat, in particular certain bourgeois ideologists who have achieved 
a theoretical understanding of the whole historical movement (14, 
pp. 87 f.). 

This organization of the proletarians as a class, and that tneans 
as a political party (14, p. 87), eventually furnishes the basis of the 
class struggle. To repeat : Every class struggle is a political struggle 
(14, p. 87). It is the deliberate and articulate conflict between two 

1 8 Marx's Model 

opposed interests, the interests, respectively, of preserving and of 
revolutionizing the existing institutions and power relations. The 
formation of classes as organized interest groups, the antagonism 
between oppressing and oppressed classes, and the resulting revolu- 
tionary changes constitute the law of development of all history up 
to now. An of -pressed class is the condition of existence of every 
society based on class conflict. Thus the liberation of the oppressed 
class necessarily involves the creation of a new society (6, p. i88). 
The history of all societies up to the present is the history of class 
struggles ( 1 4, p. 8 1 ) . 


Following these laws of development the proletariat has organ- 
ized itself in the womb of bourgeois society, and has opened its strug- 
gle against the bourgeoisie. 

Does this mean that after the downfall of the old society there 
will be a new class rule culminating in a new political authority? No. 

The condition of the liberation of the working class is the aboli- 
tion of every class y just as the condition of the liberation of the Third 
Estate, i.e., the establishment of the bourgeois order^ was the aboli- 
tion of all estates. 

The working class will in the course of development replace the 
old bourgeois society by an association which excludes classes and 
their conflict, and there will no longer be any political authority 
proper, since it is especially the political authority that provides class 
conflict within bourgeois society with its official expression. 

By now the conflict between proletariat and bourgeoisie is a strug- 
gle of one class against another, a struggle that means in its highest 
expression a total revolution. Is there any reason to be surprised 
that a society based on class conflict leads to brutal opposition, and 
in the last resort to a clash between individuals? 

Nobody should say that society develops independently of poli- 
tics. There is no political m>ovement which is not at the same time 
a social movement. 

Only in an order of things in which there are no classes and no 
class conflicts will social evolutions cease to be political revolutions 
(6, pp. i88f.). 


If Marx himself had written this chapter, it would no doubt have 
been longer, more polemical, and more directly related to the society 

Marx's Model 19 

of his time/ Nevertheless the attempt to present Marx^s approach to 
a theory of class conflict largely in his own words is more than an en- 
tertaining game. The result can serve as a fruitful basis for some 
more general observations which will prove useful for subsequent 
critical considerations. The following elements of Marx's theory of 
class appear particularly worth emphasizing for sociological analysis: 

1 . It is important to realize what Geiger called the "heuristic pur- 
pose behind the concept of class" (46, chap, ii) . Wherever Marx used 
the concept in a sociological sense, he was not concerned with describ- 
ing an existing state of society. He was concerned, rather, with the 
analysis of certain laws of social development and of the forces in- 
volved in this development. To use the misleading terms of modem 
sociology, the heuristic purpose of the concept of class was for Marx 
not "static" but "dynamic," not "descriptive" but "analytical." What 
these terms may mean and what they cannot mean will have to be dis- 
cussed later in some detail. Here it is sufficient to emphasize that for 
Marx the theory of class was not a theory of a cross section of society 
arrested in time, in particular not a theory of social stratification, but 
a tool for the explanation of changes in total societies. In elaborating 
and applying his theory of class, Marx was not guided by the question 
"How does a given society in fact look at a given point of time?" but 
by the question "How does the structure of a society change?" or, in 
his own words, "What is the [economic] law of motion of modern 
society? " 

2. This heuristic purpose explains the often criticized two-class 
model underlying the dynamic theory of Marx. Had Marx wanted 
to describe his society with photographic accuracy, this model would 
indeed have been most unsatisfactory. As a matter of fact, Marx does 
refer occasionally (without always using his concept of class in an en- 
tirely unambiguous manner) to a multitude of classes. He refers to 
the "two great categories into which the interest of the bourgeoisie 
is divided — land ownership and capital" (8, p. 43), to the petty bour- 
geoisie as a "transitional class" (8, p. 49), and to the class of small 
peasants (8, p. 118). But in principle these "intermediate and transi- 
tional stages," as Marx significantly calls them, "do not matter for 

■^ In this sense, Renner's attempt (26, pp. 374 ff.) to reconstruct this chapter of 
Capiat is closer to Marx in style and content than the attempt here undertaken to 
sketch the most general elements of Marx's theory of class. Clearly, the claim that I 
have written the unwritten last chapter of Cafital must not be understood literally 
in the sense of a philological conjecture. My main purpose has been to offer a syste- 
matic presentation of the many isolated statements about class in the work of Marx. 

20 Marx^s Model 

our investigation" (12, III, p. 941). Not only are they unstable 
entities destined to be drawn sooner or later into the two great whirl- 
pools of bourgeoisie and proletariat, but even if this were not the case, 
their historical role would be insignificant by comparison with that of 
the dominant classes of capitalist society. The concept of class is an 
analytical category, or, as Marx says in one of his rare but enlighten- 
ing methodological remarks, "real conditions are presented only in 
so far as they express their own general type" (12, III, p. 167). The 
general type of the real conditions of conflict that generates change, 
however, is the opposition of two dominant forces, two prevalent 

Geiger has refuted the unjustified objections to Marx's two-class 
model so convincingly that further discussion of them is unnecessary 
(46, pp. 37 ff.). But the legitimacy of assuming for analytical pur- 
poses the dominance of only two conflicting classes must not blind us 
to the fact that Marx has linked with his two-class model a number 
of additional postulates whose legitimacy appears rather more dubi- 
ous. For Marx the category of class defines one side of an antagonism 
which entails the dominant issues of conflict in every society as well 
as the direction of its development. This means for Marx that {a) 
every conflict capable of generating structural change is a class con- 
flict, (^) the contents of class conflict always represent the dominant 
issues of social conflict, and {c) the two classes stand in the relation 
of HegePs "thesis" and "antithesis," in the sense that one is charac- 
terized by the affirmation (or possession) of those features of which 
the other is the complete negation. It is at least open to dispute 
whether this last approach recommends itself in social science. The 
other two postulates connected with Marx's two-class model, how- 
ever, are empirical generalizations, the untenability of which will have 
to be demonstrated. Only if it is freed of these accessories can the two- 
class model be conceived as a feasible principle of knowledge. 

3. Marx has tried to argue for the third postulate mentioned 
above in the most difficult part of his theory, the part concerned with 
the causes and origins of classes. What are the structural conditions 
of the formation of social classes? For simplicity's sake I shall treat 
this aspect of Marx's theory of class with reference to his analysis of 
capitalist society, since the question remains undecided for the time 
being whether this theory can be applied to other types of society at all. 

Marx states quite clearly that class conflicts do not originate in 
differences of income, or of the sources of income. His classes are 
not tax classes in the sense of the Roman censors. Rather, the determi- 

Marx's Model 21 

nant of classes is "property." Property, however, must not be under- 
stood in terms of purely passive wealth, but as an eflFective force of 
production, as "ownership of means of production" and its denial to 
others. In this sense, the "relations of production," i.e., the authority 
relations resulting from the distribution of eflFective property in the 
realm of (industrial) production, constitute the ultimate determinant 
of the formation of classes and the development of class conflicts. The 
capitalists possess factories and machines, and buy the only property 
of the proletarians, their labor power, in order to produce a surplus 
value with these means of production and augment their capital. 

But our question cannot be answered all that easily. The role of 
property in Marx's theory of class poses a problem of interpretation, 
and on this interpretation the validity of Marx's theory of class stands 
or falls. Does Marx understand, by the relations of property or pro- 
duction, the relations of factual control and subordination in the enter- 
prises of industrial production — or merely the authority relations in 
so far as they are based on the legal title of property? Does he con- 
ceive of property in a loose (sociological) sense — i.e., in terms of the 
exclusiveness of legitimate control (in which the manager also exer- 
cises property functions) — or merely as a statutary property right in 
connection with such control? Is property for Marx a special case of 
authority — or, vice versa, authority a special case of property? These 
questions are of considerable significance. If one works with the nar- 
row concept of property, class conflict is the specific characteristic of 
a form of production which rests on the union of ownership and con- 
trol. In this case a society in which control is exercised, for example, by 
state functionaries, has by definition neither classes nor class conflicts. 
If, on the other hand, one works with the wider concept of property, 
class structure is determined by the authority structure of the enter- 
prise, and the category of class becomes at least potentially applicable 
to all "relations of production." 

Marx does not always make his answer to our questions entirely 
clear. But it can be shown that his analyses are essentially based on 
the narrow, legal concept of property. This procedure, and this pro- 
cedure only, enables Marx to link his sociology with his philosophy 
of history — a brilliant attempt, but at the same time a fault that robs 
his sociological analyses of stringency and conviction, a fault made 
no more acceptable by the fact that orthodox Marxists have remained 
faithful to their master in this point to the present day. 

The most striking evidence for this interpretation can be found in 
the preliminary attempts at an analysis of the new form of ownership 

22 Marx's Model 

characteristic of joint-stock companies which Marx presents in Volume 
III of Cafital. Marx is here explicitly concerned with the phenome- 
non that is commonly described today as the separation of ownership 
and control. He discusses what he calls the "transformation of the 
really functioning capitalist into a mere director, an administrator 
of alien capital, and of the owners of capital into mere owners, mere 
money capitalists" (12, III, p. 477). "In joint-stock companies, func- 
tion is separated from capital ownership j thereby labor is entirely sep- 
arated from ownership of the means of production, and of surplus 
labor" (p. 478). Now, hard though it is for ordinary minds to see 
why this change in the size and legal structure of industrial enter- 
prises should end the conflict between entrepreneurs who can com- 
mand and workers who have to obey (the conflict that Marx postulates 
for the "pure" capitalist enterprise), Marx ascribes to the joint-stock 
company a peculiar place in history. Time and again he describes the 
joint-stock company as "private production without the control of 
private property" (p. 480), as "the elimination of capital as private 
property within the capitalist mode of production itself" (p. 477), 
and even as the "abolition of the capitalist mode of production within 
the capitalist mode of production itself " (p. 479). Forhim, the joint- 
stock company is "a necessary point on the way to reconverting capital 
into the property of the producers, this no longer being the private 
property of individual producers but their associated property, i.e., 
immediate social property" (p. 478). It is a "point on the way to the 
transformation of all functions in the process of reproduction hitherto 
connected with capital ownership into mere functions of the associated 
producers, into social functions" (p. 478). The joint-stock company, 
in other words, is halfway to the communist — and that means class- 
less — society. 

We cannot pursue here the manifold consequences of this strange 
analysis, which — correct as it may to a certain extent be empirically 
(if in a sense hardly intended by Marx) — would certainly have ex- 
posed Marx, had he lived longer, to many an awkward question from 
his most orthodox adherents. One point, however, is convincingly 
demonstrated by this analysis: for Marx, the relations of production 
as a determinant of class formation were also authority relations, but 
they were such only because in the first place they were property rela- 
tions in the narrow sense of the distribution of controlling private 
ownership. Qua property relations they are authority relations, and 
not vice versa, not qua authority relations property relations. If, 
therefore, the functions of the "director" and the "mere owner," the 

Marx^s Model 23 

manager and the stockholder, are separated, this means a first step on 
the way to the complete abolition not only of effective private prop- 
erty itself, but also of the authority relations dependent on it, and thus 
a step on the way to the communist society. For Marx, classes were 
tied to the existence of effective private property. Their formation, 
existence, and struggle can occur only in a society in which some possess 
and others are excluded from private ownership and control of the 
means of production. 

4. One of the critical pivots of Marx^s theory of class is the undis- 
puted identification of economic and political power and authority. 
Although classes are founded on the "relations of production," i.e., 
the distribution of effective property in the narrow sphere of com- 
modity production, they become socially significant only in the politi- 
cal sphere. But both these spheres are inseparable. "The political 
power" of a class arises for Marx "from the relations of production" 
(5, p. 455). The relations of production are "the final secret, the 
hidden basis of the whole construction of society" ( 1 2, III, p. 842 ) •■, 
industrial classes are eo ipso also social classes, and industrial class 
conflict is political class conflict. Nowhere has Marx explicitly dis- 
cussed the basis of this empirical proposition — nor has he seen suf- 
ficiently clearly that it is an empirical proposition rather than a postu- 
late or premise. The thesis that political conditions are determined 
by industrial conditions seems to stem, for him, from the generalized 
assertion of an absolute and universal primacy of production over all 
other structures of economy and society. It is evident that a postulate 
of this kind requires empirical testj how it fares in this test will have 
to be shown. 

5. Relatively thoroughly, if nowhere systematically, Marx has 
described the steps of the process by which groupings in the form of 
classes emerge from conditions of social structure. For Marx, the 
first stage of this process of the formation of classes is given directly 
by the distribution of effective private property. Possession and non- 
possession of effective private property create two peculiar "common 
situations," "conditions of life," or class situations. These class situa- 
tions have three complementary aspects: {a) that of the mere distri- 
bution of effective property, i.e., of possession or nonpossession of 
means of production and of authority 5 {b) that of the possession or 
nonpossession of goods and values gratifying personal needs, i.e., the 
"rewards" of modern sociology; and {c) that of the common situa- 
tionally determined interests of those who share a class situation. By 
common interest in Marx's sense is not meant a conscious tendency 

24 Marx's Model 

of individual desires, but a potentially unconscious (or "falsely con- 
scious") tendency of actual behavior shared by people in a common 
class situation. Common interests exist, as Marx says, "not merely 
in the imagination, . . . but above all in reality as the mutual de- 
pendence of the individuals among whom labor is divided" (13, II, 
p. 23). This is a difficult notion j for we are used to conceiving of 
interests above all on a psychological level. For the time being, how- 
ever, we shall put off considering in what sense a concept of "objective 
interests" that "exist as" real conditions may be useful. At this point 
we merely conclude that Marx's theory of class formation starts with 
the postulate of a common class situation, the main components of 
which are a common relation to effective private property, a common 
socio-economic situation, and a common tendency of actual behavior 
determined by "objective" interests. 

In accordance with the premises of the theory of class, we already 
find, from this point of view, a fundamental dichotomy of class situa- 
tions in any given society, and of the members of a society by their class 
situations. Occasionally, Marx refers to the aggregates thus defined 
as classes. "In so far as millions of families live under economic con- 
ditions which separate their way of life, their interests, and their edu- 
cation from those of other classes and oppose them to these, they con- 
stitute a class" (8, p. 104). The concept of class as defined so far cor- 
responds to Max Weber's later formulation, " 'Class' shall mean any 
group of persons in a common class situation" (33«, p. 177).* How- 
ever, this definition has its problems. It has to be asked whether a 
common situation is sufficient to constitute a group in the strict sense 
of this term. If — as can be shown — this is not the case, it remains to 
be asked how an aggregate of people who are merely situated identi- 
cally, without having any contact or coherence, can become an effective 
force in social conflict and change. Marx has asked this question, and 
he therefore emphasizes at many points that the mere "gap between 
the conditions of life," the mere "identity of interests" and class situa- 
tions, is a necessary but by no means sufficient condition of the forma- 
tion of classes. He continues therefore in the passage quoted above: 
"In so far as there is a merely local contact [among people in a com- 
mon class situation] — in so far as the identity of their interests does 

^ As a matter of fact, Weber is well aware of the problem under discussion. He 
therefore distinguishes "property classes" (as "not 'dynamic' ") from "income classes" 
and "social classes." However, since Weber describes all three classes as consisting of 
all people in a common (if differently defined) class situation, his theory of class lacks 
the analytical strength of Marx's, which is much more precise on this point. 

Marx*s Model 2$ 

not produce a community, national association, and political organi- 
zation — they do not constitute a class. They are therefore unable to 
make their class interest heard. . . ."(8, p. 104). 

6. This may well be the most important step in Marx's theory 
of class formation: Classes do not constitute themselves as such until 
they participate in political conflicts as organized groups. Although 
Marx occasionally uses the concept of class in a less determinate, more 
comprehensive sense, a multitude of statements leave little doubt that 
for him class formation and class conflict were phenomena belonging 
to the sphere of politics. "As long as the proletariat has not sufficiently 
developed to organize itself as a class, . . . the struggle of the pro- 
letariat with the bourgeoisie has not yet assumed a political character" 
(6, p. 142). This means conversely that the carriers of class conflict 
have organized themselves as classes, and have become classes, only 
if class conflict has assumed a political character. 

For Marx, this last stage of class formation has two complemen- 
tary aspects. On the factual level of social structure it involves the 
association of people who share a class situation in a strict group, party, 
or political organization. Marx refers to the "organization of the 
proletarians as a class, and that means as a political party" (14, p. 87). 
On the normative and ideological level of social structure it involves 
the articulation of "class-consciousness," i.e., the transformation of 
"objective" class interests into subjectively conscious, formulated 
goals of organized action. The complete class is characterized not by a 
common though unconscious direction of behavior, but by its conscious 
action toward formulated goals. 

7. Marx's theory of class formation is embedded in his work in 
a wider theory of class conflict as the moving force of social change. 
However, the elements of this wider theory are only partly of a so- 
ciological nature. It contains a number of theses whose validity can no 
longer be tested by empirical research. In summarizing those ele- 
ments of Marx's theory of class conflict that are of potential use to the 
sociologist, we soon reach the point at which Marx the sociologist and 
Marx the philosopher joined forces: 

{a) In every society there is possession of and exclusion from 
effective private property. In every society there is therefore posses- 
sion of and exclusion from legitimate power. The "relations of pro- 
duction" determine different class situations in the sense indicated 

{b) Differentiation of class situations toward the extremes of 

26 Marx's Model 

possession of and exclusion from property and power increases as a 
society develops. 

{c) As the gap between class situations grows, the conditions of 
class formation — i.e., of political organization and of the explicit 
formulation of class Interests — mature. The political class struggle 
between "oppressors" and "oppressed" begins. 

{d) At its climax this conflict produces a revolutionary change, in 
which the hitherto ruling class loses Its power position and is replaced 
by the hitherto oppressed class. A new society emerges. In which a new 
oppressed class grows up, and the process of class formation and class 
conflict starts anew. 

It will prove necessary to subject this wider theory of Marx to 
severe criticism from a sociological point of view,^ even though the 
most problematical aspects of Marx^s theory, such as the notion of a 
classless society, have so far been left out of consideration. Here we 
are concerned merely with a resume of the sociological elements of 
Marx's theory of class. It is, as we can say now, a theory of structural 
change by revolutions based on conflicts between antagonistic interest 
groups. Marx describes in detail the structural conditions and the 
process of formation of these interest groups. Less elaborately, but 
clearly enough, he also describes the process of conflict between these 
groups and its solution in revolutionary changes. 

8. Before we leave Marx's sociology one more formal character- 
istic of his theory of class warrants recognition, since it Is not without 
significance for recent sociological theory. By analyzing the change of 
social structures In terms of the categories mentioned, Marx Intro- 
duces at least implicitly a certain image of society. Although an image 
of society of this kind may not be of immediate empirical relevance 
for sociological research. It can nevertheless become a measure of the 
proximity of a theoretical construction to reality, and it serves im- 
portant functions as a guide to problems of research.^" 

® This holds in particular for propositions {a) and {b), which can only be em- 
pirical generalizations and are as such untenable even if Marx abandons them arbi- 
trarily for the two societies he invented: the "original society" and the "final society" 
of history. Proposition (c) is also problematical; see the section on "class conflict and 
revolution" in Chapter IV. Generally speaking, it is my intention in this chapter — 
at least in so far as sociological questions are concerned — to indicate the points of 
departure of criticism, but to postpone the criticism itself. 

^° See Chapter III, pp. 112 ff., and Chapter V, pp. 157 ff., for a more elaborate 
discussion of "images of society." 

Marx's Model l-j 

For Marx, society is not primarily a smoothly functioning order 
of the form of a social organism, a social system, or a static social 
fabric. Its dominant characteristic is, rather, the continuous change 
of not only its elements, but its very structural form. This change in 
turn bears witness to the presence of conflicts as an essential feature 
of every society. Conflicts are not random j they are a systematic prod- 
uct of the structure of society itself. According to this image, there is 
no order except in the regularity of change. "Without conflict no 
progress: this is the law which civilization has followed to the present 
day" (6, p. 80)." 

This image of society stands in clear contradiction to the images 
which lie at the basis of the considerations of some recent sociologists. 
At the same time, it appears considerably more useful for the solution 
of many problems of sociological analysis than all analogies, explicit or 
implicit, between society and organism, or society and one or another 
(essentially "closed") functional system. The reality of society is 
conflict and flux. Despite our radical criticism of Marx's theory of 
class, this implication may therefore be retained as a fruitful heuristic 

In the imaginary chapter of Capital above, I have deliberately 
emphasized the nonsociological elements of Marx's theory of class 
rather less than Marx himself would have done. Now, too, I shall 
discuss them less elaborately than the sociological elements, since this 
study is intended not as a philosophical discussion or merely as a criti- 
cism of Marx, but rather as a means of posing anew the problem of 
social conflict and its sociological analysis, and furnishing some ele- 
ments of its solution. Nevertheless, it would hardly be justifiable to 
represent Marx's theory of class and subject it to critical examination 
without mentioning its nonsociological elements. 

A word of explanation may be in place as to the identification of 
its nonsociological elements with its "philosophical" elements. It 
seems to me that it can be shown that the work of Marx falls into two 

^^ Unfortunately, the clause "to the present day" is meant to imply that one day 
this law will no longer hold. Here as elsewhere Marx has vitiated the value of his 
sociology by Hegelian philosophical additions of litüe plausibility. The image of 
society indicated in the last paragraph is for Marx an image of historical societies in 
the period of alienation. Communist society (as well as the early communal society 
of Marx's philosophical imagination) is different, and indeed in many ways not dis- 
similar to the constructions of modern sociological theory. 

28 Marx's Model 

separable parts (c£. 225). On the one hand, there are categories, 
hypotheses, and theories which permit empirical test, i.e., which either 
can be falsified themselves by empirical observations, or allow of 
derivations that can be so falsified. This is true, for example, of the 
proposition that structural change is a result of class conflicts. If I 
use the term "sociological" for such elements of the work of Marx, 
I am well aware that it is too narrow with respect to Marx^s strictly 
economic propositions and theories. In the case of the theory of class, 
however, it is undoubtedly applicable. 

On the other hand, the work of Marx contains postulates and 
theories utterly removed from the possibility of empirical test. Propo- 
sitions such as that capitalist society is the last class society of history, 
or that communist society leads to a complete realization of human 
freedom, can be disputed and denied, but they cannot be refuted with 
the tools of science. To be more precise, it is impossible to imagine 
empirical data which would falsify these postulates or their deriva- 
tions. For such assertions, which are irrefutable in principle, I use the 
term "philosophical." This expression is evidently as little exhaustive 
in terms of the content and method of philosophy as is the term "socio- 
logical" in the meaning here proposed for sociology. The attributes 
"sociological" and "philosophical" signify in this context a difference 
in the logical status of propositions. Marx's theory of class contains 
elements of both kinds. Indeed, nowhere has Marx linked both kinds 
of propositions as cleverly, and hence as deceptively, as in his theory 
of class. 

Marx's conception of the communist society, of its role in history, 
and of the time of its arrival is the pivotal point of the connection be- 
tween the philosophical and the sociological elements of his theory 
of class. Later, we shall have to consider sense and nonsense in the 
notion of a classless society from a sociological point of view. Here we 
are concerned with the place of the classless society in Marx's philoso- 
phy of history. In an earlier study I have tried to show in detail in 
what sense the historical process is for Marx a dialectical process of 
thesis, antithesis, and synthesis (224). Not only does this hold for 
what is sometimes called the "real dialectics" of history from antique 
to feudal and further to capitalist society, but it holds above all for 
the total process of history. For Marx, the supreme meaning and law 
of historical development lies in the birth of human freedom, or of 
free man by human labor. In its beginning there is the postulated 
reality of an original society, in which man is "with himself" and free, 
though as yet only in a restrained and partial fashion. Private prop- 

Marx^s Model 29 

erty, classes and class conflict, division of labor, and inequality are 
absent from this society. But man is still quasi-unborn j he has not 
developed his talents yet, not recognized his potentialities j his free- 
dom is a dull freedom, without conscious activity and constrained 
by extraneous purposes. This original society breaks up with the ap- 
pearance of private property and the division of labor, and is trans- 
formed into its antithesis. 

Then begins a second stage of human development, which com- 
prises all known history. In this stage, man is alienated from himself. 
Division of labor and private property create relations of domination 
and subjection, class formations, and class struggles in ever-changing 
patterns. Indeed, private property is the specific difference of this 
second stage of the historical process, as it had been the principle of 
social inequality and social organization for Locke and Hume, Fergu- 
son and Millar. In this stage man is unfreej but the growth of the 
forces of production leads him to develop ever new potentialities and 
talents. He cannot develop these talents to the full so long as he is 
alienated and enslaved, but they are present now within him. For all 
men to develop all their talents in all directions by purpose-free ac- 
tivity, all that is needed is regained liberty, de-alienation. 

This is realized by the synthesis of classless society. It is classless, 
because in it there is no private property, and there are therefore — 
by virtue of Marx's (false) conception of property — no relations of 
authority, no class conflicts. It is a synthesis, because it combines the 
dull liberty of original society and the differentiated human poten- 
tialities of alienation in the triple sense of the Hegelian concept; it 
abolishes them as such and yet preserves them on a higher level." In 
this society man realizes himself as a free being. 

Thus formulated, Marx's philosophy of history is little more than 
Hegel's theory of history as the realization of the spirit of freedom 
*'turned upside down." In this formulation also, Marx's philosophy 
of history is not directly relevant to the sociological theory of class. 
However, Marx joins the two by a fascinating trick of definition and 
thereby manages to give his philosophy the appearance of empirical 
validity, and his sociology the force of indubitable truth. As long as 
these two logical bastards were not recognized as such, they could 
have, singly and combined, that political effect so well known today. 

■"^ The word Hegel and Marx use most frequently to characterize the synthesis 
of the dialectical process is aujheben. Even as a word, it has three meanings, all of 
which were intended by Hegel and Marx: (l) to suspend or abolish, (2) to elevate or 
lift to a higher level, (3) to preserve or maintain. 

30 Marx^s Model 

It might seem that the concept of class is not really necessary for 
Marx^s philosophy of history. In speaking of the transition from the 
original society to the state of alienation, Marx sees the cause of this 
"fall of man" in private property. Correspondingly, the central event 
of the transition from the state of alienation to the realm of liberty 
is the abolition of efiFective private property. All other symptoms of 
alienation, such as classes and power relations, the state and the divi- 
sion of labor,^^ also disappear. But they almost look like empirical 
trimmings with no particular relation to the argument — unless they 
are asserted to be dependent on private property. 

In a logically independent approach Marx investigates the society 
of his time. There he notes empirically three factors, among others: 
{a) the presence of a conflict between social groups (classes), {b) the 
presence of efiFective private property, and {c) the presence of rela- 
tions of domination and subjection. He believes furthermore that he 
can discern in this society a tendency for private property to be abol- 
ished and replaced by communal property — an observation which at 
least to some extent has proved correct. And what happens if efiFective 
private property disappears? It is precisely at this point that Marx 
jumps from sociology into philosophy and back by introducing his 
undoubtedly brilliant trick of definition. By asserting the dependence 
of classes on relations of domination and subjection, and the depend- 
ence of these relations on the possession of or exclusion from efiFective 
private capital,^* he makes on the one hand empirically private prop- 
erty, on the other hand philosophically social classes, the central fac- 
tor of his analyses. One can retrace step by step the thought process 
to which Marx has succumbed at this point. It is not the thought 
process of the empirical scientist who seeks only piecemeal knowledge 
and expects only piecemeal progress, but that of the system builder 
who suddenly finds that everything fits! For if private property dis- 
appears (empirical hypothesis), then there are no longer classes (trick 
of definition) ! If there are no longer any classes, there is no alienation 
(speculative postulate). The realm of liberty is realized on earth 

^^ Marx's considerations are least unambiguous in this question of the division 
of labor. Thus when Marx and Engels try to prove in their German Ideology that 
there will be no division of labor in communist society, their proof remains under- 
standably unconvincing and is really confined to the thesis that the final society will 
replace the "specialist," or "detail man," by the "universal man." 

■^* An assertion which is of course understandable in view of the factual identity 
of ownership and control in early industrial capitalism, but which is unpardonable as 
a generalization. Cf. pp. 41 ff. 

Marx's Model 3 1 

(philosophical idea). Had Marx, conversely, defined private prop- 
erty by authority relations, his empirical observation would not have 
"fitted," and he would have had to drop his philosophy of history. 
For effective private property may disappear empirically, but au- 
thority relations can do so only by the magic trick of the system 

There would be no objection to joining philosophy and sociology, 
if this were done without consequences detrimental to sociology. This 
is not, however, the case here. Marx's philosophy has forced him to 
betray his sociology, and this betrayal forces us to separate the two 
elements relentlessly. Far-reaching as this assertion may sound, it 
can easily be justified by reference to Marx's work, as a few examples 
may show: 

1 . The dogmatic conj unction of classes and effective private prop- 
erty documents in itself a betrayal of sociology. Perhaps a Marx with- 
out the Marxian philosophy of history would have realized that power 
and authority are not tied to the legal title of property. Marx himself 
could not realize this, and certainly could not admit it, for had he done 
so, his philosophical conception of the classless society would have 
become impossible both empirically and intellectually. In other words, 
he had either to regard the joint-stock company as a transitional form 
on the way to the classless society, or to abandon the philosophy ac- 
cording to which private property — and not the possession of or exclu- 
sion from authority — is the differentia sfecifca of alienation and the 
determining factor of classes. 

2. In asserting the universality of class conflict in the stage of 
alienation, Marx has again saved his philosophy and sacrificed his 
sociology. "The history of all societies up to the present is the history 
of class struggles." This seemingly empirical sentence is in reality 
but a reformulation of the philosophical postulate that links alienation 
(and thereby all known history), private property, and the classes. It 
becomes a dogma that prevents open, unprejudiced research from 
developing. There is, moreover, a strange irony in the fact that the 
same Marx who so often attacked the uncritical assertion that private 
property is universal, introduces the same assertion in a concealed way 
but equally uncritically by speaking of the universality of classes, 
which for him are tied to the presence of private property. 

3. Marx's analysis of his own capitalist society is evidently col- 

^° The thesis of this argument was intimated first by Schumpeter. Cf. 73, pp. 19 f. 
See also below, pp. 84 flF. 

32 Marx's Model 

ored by his sociologically nonsensical conviction that it is the last class 
society in history. Thus his assertions of the extraordinary intensity 
of class conflict in this society and of the messianic role of the prole- 
tariat have little to do with empirical knowledge. They invest the 
concept of class with an importance and a measure of exclusiveness in 
sociological analysis which is clearly not warranted by historical ex- 
perience (although it has blinded generations of sociologists). 

4. Finally, the application of dialectics as the asserted inherent 
law of historical development involves a betrayal of sociology in the 
interests of philosophy. This is especially evident in the misleading 
and sociologically untenable thesis that structural change is necessarily 
of a revolutionary character — once again a thesis unfortunately im- 
pressive enough to divert the eyes of social scientists away from real 

Several further points of lesser importance could be mentioned 
at which sociological and philosophical elements are linked in Marx's 
theory of class. However, such points are not our concern here. We 
are concerned rather with a rough sketch of the nonsociological as- 
pects of Marx's theory of class, and at the same time with finding a 
point of departure for their radical separation from the sociological 
aspects. For the sociologist, there can be no doubt that this separation 
is necessary. But it might be asked: Why all this effort? Why not a 
completely new and more fruitful approach? There are two answers 
to these questions: First, it is often useful to start with a critical review 
of the errors of earlier authors so that one can avoid, them oneself. 
Second, Marx's theory of class, if freed of all speculative trimmings, 
contains many insights and useful approaches, which we can scarcely 
afford to ignore. 

Our discussion has brought out the main lines of Marx's view of 
the society of his time and its development. To this view some details 
may be added now which can usefully serve as a point of departure for 
the analysis of factual changes in industrial societies since Marx, in 
so far as they relate to the class structure of these societies. Following 
Marx, we shall use the term "capitalist society" without defining it 
precisely for the time being. But unlike Marx, we shall ignore the 
philosophical elements of his analysis and renounce the cheap triumph 
of "refuting" speculative prophecies by empirical data. 

According to Marx, capitalist society is a class society. There is 
in this society a category of persons who possess effective private prop- 

Marx's Model 33 

erty, and another category of those who have no such property. The 
former is called capital or bourgeoisie, the latter wage labor or pro- 
letariat. The typical private property of capitalist society consists of 
the means of industrial production, i.e., factories, machines, and the 
like, or capital. The owners or capitalists directly control their means 
of production j the nonowners or wage laborers are dependent, by the 
labor contract, on the means of production and their owners. Property 
and power and the exclusion from both go together; they "correlate." 
There is also a correlation between these factors on the one hand, and 
socio-economic position on the other hand: the capitalists are wealthy, 
secure, and have high status j the wage laborers are lacking a sub- 
sistence minimum. This difference in position makes for conflicting 
interests and conflicting groupings — classes — which fight each other 
at first on the local level of the individual enterprise, eventually on 
the political level. 

There are, of course, persons in capitalist society — such as land- 
lords, independent craftsmen and small businessmen, peasants, and 
intellectuals — who stand outside this tension and whose interests are 
not directly afFected by it. These groups, however, not only decrease 
in numerical importance but increasingly lose their influence on the 
conflicts determining the structure of society. The capitalist bour- 
geoisie and its counterpart, the industrial proletariat, move more and 
more into the center of the social process. Their conflicts dominate the 
scene of capitalist class society and draw all other groups into their 
orbit or condemn them to complete insignificance. Society is domi- 
nated by the antagonism between the interests of those who defend 
their possession of eff^ective private property and those who elevate 
their nonpossession into a demand for a complete change of the prop- 
erty relations. 

This sketch of Marx^s view of capitalist society is incomplete in 
one important point. It describes a structure, and not its process of 
development, whereas it is on the latter count that Marx made his 
sociologically important contribution to social analysis. Marx tried, 
at times retrospectively, more often predictively, to determine the 
tendencies of change that can be derived from this structure. With 
respect to the development of class structure he emphasized in particu- 
lar the following four processes : 

I. Inherent in capitalist society, there is a tendency for the classes 
to polarize increasingly. "The whole society breaks up more and more 
into two great hostile camps, two great, directly antagonistic classes: 
bourgeoisie and proletariat" ( 14, p. 7). Here, the model of two domi- 

34 Marx*s Model 

nant classes is no longer merely a heuristic postulate, but describes a 
factual condition. "The earlier petty bourgeoisie, the small indus- 
trialists, the merchants and rentiers, the craftsmen and peasants, all 
these classes sink down into the proletariat" (14, p. 16). It is really 
misleading to speak of two "great" classes, since social development, 
according to Marx, produces a polarized class society with a relatively 
small ruling class of capitalists and an extraordinarily large oppressed 
class of wage laborers. 

2. As the classes polarize, their class situations become increas- 
ingly extreme. On the one hand, the wealth of the bourgeoisie is 
swelled by larger profits based on increasing productivity as well as 
by the progressive concentration of capital in the hands of a few indi- 
viduals. "One capitalist kills many others" (12, 1, p. 803). On the 
other hand, "with the continuously decreasing number of capital mag- 
nates who usurp and monopolize all advantages of this process of 
change . . . [comes an increase in] the mass of poverty, of pres- 
sure, of slavery, of perversion, of exploitation, but also of revolt on 
the part of a working class permanently increasing in size which is 
skilled, united, and organized by the mechanism of the capitalist mode 
of production itself" (12, I, p. 803). Here, the so-called theory of 
pauperization has its place, according to which the poverty of the pro- 
letariat grows with the expansion of production by virtue of a law 
postulated as inherent in a capitalist economy. 

3. At the same time, the two classes become more and more ho- 
mogeneous internally. In the beginning of this process, the classes 
are clearly delimited from the outside, but rather heterogeneous 
within. Marx says of the bourgeoisie that its members "have identical 
interests in so far as they form a class in opposition to another class," 
but "contradictory and conflicting interests as soon as they are con- 
fronted with themselves" (6, p. 140). Analogously, the proletariat 
is not, in the beginning, a "class for itself." However, a number 
of processes mold the different constituents of the classes into uniform 
groups without significant internal differences or conflicts. "More 
and more the collisions between individual workers and bourgeois 
assume the character of collisions of two classes" (14, p. 87). This 
is partly due to pressure from without, such as the growing intensity 
of the class struggle. Partly it is the effect of social and even technical 
factors. In the case of the proletariat Marx refers on the one hand to 
the growing extent of class organization as a unifying factor, on the 
other hand to the "tendency of equalizing and leveling in work proc- 
esses" within industry itself (12, I, p. 441), i.e., the reduction of all 

Marx's Model 35 

workers to unskilled laborers by the technical development of produc- 
tion. Similarly, a combination of economic, and in the narrow sense 
social, factors unites the bourgeoisie as a class. 

4. Once history has carried these tendencies of development to 
their extremes, the point is reached at which the fabric of the existing 
social structure breaks and a revolution terminates capitalist society. 
The hitherto oppressed proletariat assumes power j efFective private 
property is socialized j classes cease to exist j the state is withering 
away. The proletarian revolution inaugurates the communist, class- 
less society. 

Marx's image of capitalist society is the image of a society under- 
going a process of radical change. This change culminates in a revo- 
lutionary act, into which all earlier developments converge and from 
which all later developments depart. The executors of this process are 
structurally generated, organized human interest groups — ^the classes. 
One of these — the bourgeoisie — defends with sinking chances of 
success the existing distribution of property, and with it the whole 
social status quo. The other one — the proletariat — attacks this status 
quo with growing success until the day on which its interests become 
reality, the values of a new society. The capitalist form of economic 
and social structure is doomed, and the classes are its gravediggers. 

At this point in our considerations we shall depart from Marx. 
His theory of class provides the background of subsequent argument, 
his analysis of capitalist class society that of later analyses. If we suc- 
ceed in refuting the sociological theories of Marx or the hypotheses 
derived from them, we have good reason to rejoice. For science grows 
by the refutation of accepted propositions and theories, and not by 
their stubborn retention. 


Changes in the Structure of Industrial Societies 

Since Marx 


It is a commonplace today that many of Marx's predictions have 
been refuted by the social development of industrial societies over 
the past century. But while commonplaces of this kind can be an 
excuse for silence, they are also a challenge. If a commonplace is 
repeated too often, it makes one wonder whether those who repeat it 
realize what they are saying and are able to substantiate their asser- 
tions. In sociological literature, there has been, to my knowledge, but 
one attempt to present a systematic account of the social changes that 
molded industrial societies in forms unforeseen by Marx, namely, 
Theodor Geiger's analysis of the Class Society in the Melting-Pot. 
This account, while useful, is rather less than exhaustive, to say noth- 
ing of the fact that Geiger fails to present these changes with a view 
to formulating a new and better theory of social conflict and change 
than was provided by Marx.^ If, therefore, I propose to explore in 
this chapter some of the important changes in the structure of indus- 
trial societies since Marx, I do so with a double purpose. First, I want 
to indicate certain patterns of social development that justify speaking 
of Marx's theory of class as being falsified by empirical observations. 
Second, however, I intend to discuss those features of advanced in- 
dustrial societies which have to be accounted for by a theory of con- 
flict and change that claims to be applicable not merely to capitalist 
societies but to industrial societies in general. For both these purposes, 
and in the context of a study that is largely analytical in nature, it is 
not necessary to waste much time and space in recounting well-known 
facts of development. Accordingly, this chapter, although based on 
observations and empirical generalizations, will be devoted to analyses 
rather than descriptions of historical trends, and will thus clear the 
ground for our further considerations. 

■^ For an extensive discussion of Geiger's book see below, Chapter III, pp. 97 S. 

Changes Since Marx 37 

Marx dealt with what he called "bourgeois" or "capitalist" so- 
ciety. By contrast, the title of this study and of this chapter refers to 
"industrial society." In describing modern societies, social scientists 
have tended to use these concepts either indiscriminately or confus- 
ingly. There are several reasons for this confusion. "Whether con- 
temporary Britain, or contemporary America, can properly be de- 
scribed as capitalist societies is partly a question of fact and partly a 
question of terminology" (T. Bottomore, 37, p. 13 n). In so far as 
it is a question of terminology, i.e., of an essentially arbitrary decision, 
political rather than scientific motivations play their part. Some au- 
thors — including Bottomore (37) and H. P. Bahrdt ( 1 24) — feel that 
in preferring the concept of industrial society, sociologists have aban- 
doned the impulses of critical detachment from social reality which 
stood at the origin of social science, and which are even today conveyed 
by the notion of capitalist society. Those who speak of industrial so- 
cieties, it is claimed, have accepted the existing state of affairs as un- 
objectionable in a way in which those who prefer to speak of capitalist 
societies have not. There is an element of truth in this allegation. 
Sociologists cannot be said to be free of the suspicion of ideology be- 
cause they investigate, inter alia, such suspicions. However, despite 
this element of truth I propose to ignore the charge of an ideological 
bias related to the concepts here in question and concentrate on what 
Bottomore calls the "question of fact" involved in the terminological 
dispute. For it seems to me that despite some evidence to the con- 
trary the notions of "capitalist society" and "industrial society" are 
not different terms for identical concepts, but in fact signify different 
concepts which we have to disentangle in order to clarify the logical 
and historical status of our analyses. 

"Capitalism" was originally — and still is to some extent — an eco- 
nomic concept. The notion of a capitalist society is an extrapolation 
from economic to social relations j it assumes some formative power 
on the part of economic structures if not indeed the thesis that social 
institutions and values are but a superstructure on the real basis of 
economic conditions. However, not even economists wholly agreed, 
or agree, on the characteristics of this "real basis," to say nothing of 
its social "superstructure." In describing the social and economic 
order emerging in consequence of the industrial revolution in the late 
eighteenth and early nineteenth centuries, Marx emphasized private 
ownership of the means of production, the "free" sale of labor, the 
production of surplus value, mechanized factory production, and the 
existence of classes. Sombart and Max Weber later added the rational 

3 8 Changes Since Marx 

or rationalist value system, the acquisitive principle, the market char- 
acter of the economy. According to Weber, it is "one of the funda- 
mental properties of capitalist private economy that it rationalizes on 
the basis of strict calculation, and that it is structured systematically 
and soberly by the economic effect intended" ( 1 89, p. 61 ) . With de- 
ceptive thoroughness, Sombart defines capitalism as "a commercial 
organization in which two different groups of the population are con- 
nected by the market and cooperate regularly: the owners of the 
means of production, who at the same time have control and are the 
subjects of the economy, and propertyless mere laborers as objects of 
the economy J it is an organization which is dominated by the acquisi- 
tive principle and by economic rationalism" (28, 1, 319). Many fur- 
ther "definitions" of capitalism could be quoted here, but they add 
little to those of Marx, Weber, and Sombart.^ 

On the other hand, it is evident by definition that — in Heimann's 
words — "capitalism is a historical concept" : "It signifies one economic 
system among other systems realized in history, and it therefore im- 
plies the problem of the origin and, possibly, the transformation and 
supersedure of capitalism" (177, p. 510). Thus, the derived concept 
of a capitalist society also signifies a historical pattern of social struc- 
ture that can change and be replaced by new patterns. Here the ques- 
tion arises which makes all ambiguities of definition apparent: What 
must change for a capitalist system to cease being capitalist? 

Let us take, for example, Sombart^s definition quoted above. It 
contains at least seven elements or determinants of capitalism: (i) 
the commercial organization, (2) the cooperation of two groups of the 
population, (3) the fact that one of these groups simultaneously owns 
and controls the means of production, whereas (4) the other group 
has no property and is confined to "merely" laboring, (5) their con- 
nection by the market, (6) the acquisitive principle, and (7) economic 
rationalism. What happens — we must ask — if one of these factors 
changes, but all others remain unchanged? What happens, for ex- 
ample, if ownership and control of the means of production are no 

^ Apart from the empirical problem, all these "definitions" also raise a logical 
problem: Where, in social science, do definitions end and descriptions begin? A 
more elaborate discussion of this problem would reveal some of the darkest corners of 
social science. In this study, I shall speak of definitions only with respect to the defi- 
nition of concepts in terms of their genus and dijferentia sfecißca, and I shall dis- 
tinguish them as such stricdy from statements about objects thus conceptually defined, 
which will be referred to at times as "definitions" by contrast to definitions proper. 
What matters in a definition is not that it is "complete" (which is in fact the trouble 
with Sombart's "definition" quoted above) but that it is precise and unambiguous. 

Changes Since Marx 39 

longer simultaneous, but everything else remains as before? Are we 
then still dealing with a capitalist system or not? 

If, like only too many others, we accept the given definition as in 
some sense "true," then there are two equally nonsensical answers to 
this question: either a change of one element of the "definition" is 
seen as "necessarily" involving changes in all the others, so that capi- 
talism is superseded j or a change of one element is neglected as a mere 
modification, and the concept of capitalism is applied also to systems 
which do not fulfill all conditions of the "definition." The first of 
these is Marx^s answer. It presupposes a whole series of untenable 
postulates and becomes a mere trick of definition which may deceive 
by its brilliance, but which cannot convince. The second and more 
frequent answer results in an extension of the term "capitalism" be- 
yond the borderline of meaning. Its "definition" is no longer a defi- 
nition, for it is lacking a differentia specifica. One could carry this 
procedure ad absurdum by pointing out the limiting case in which the 
concept of capitalism becomes identical with the genus economic and 
social system. There can be little doubt that neither method is very 
fruitful for historical and sociological analysis. 

For the concept of a capitalist economy or society to be useful, it 
is necessary to define it specifically and precisely. If we try to do so, 
a significant fact becomes apparent. The elements of most of the 
traditional definitions of capitalism fall into two distinct groups. On 
the one hand, we encounter factors which can be shown to be connected 
with industrial production as such, independent of its social, legal, or 
economic context: for example, the participation in production of a 
controlling group and a subordinate group, economic rationalism, pos- 
sibly some form of a market economy, and other factors to be deter- 
mined presently. On the other hand, there appear in these definitions 
elements which characterize merely the particular form of industrial 
production displayed by the industrializing countries of Europe and 
North America in the nineteenth century: above all the union of 
ownership and control, but also the poverty of industrial workers, the 
profit motive, and some other features. It is true that these two groups 
of factors occurred together in the particular situation of Europe and 
the United States in the nineteenth century. But after the experience 
of other types of industrialization we know today that their combina- 
tion was, from the point of view of industrialization and its social im- 
plications, an accident, or rather the result of certain very special his- 
torical configurations. If some social scientists claimed, and still claim, 
that there is a necessary connection between both groups of factors, 

40 Changes Since Marx 

and that therefore both will disappear at the same time, they are ill- 
advised j and we shall have to revise this position. 

Those factors which can be shown to be generated by the structure 
of industrial production, and which cannot disappear, therefore, unless 
industry itself disappears, will be associated in this study with the con- 
cept of industrial society. Since it appears evident that industrial pro- 
duction is not just a passing guest in history, but will probably be with 
us forever in one form or another, it follows that the concept of an 
industrial society is extremely comprehensive. Whenever it is applied 
to particular societies it will require specification. In general, how- 
ever, we retain mechanized commodity production in factories and 
enterprises as the distinguishing feature of industrial societies. The 
social conditions that correlate with this characteristic can be specified 
only by way of empirical analysis. 

By contrast, capitalism merely signifies one form of industrial so- 
ciety. Definitions of terms always contain an element of arbitrariness, 
and one could certainly define capitalism differently. However, it 
accords with the definitions of the most acute modern economists to 
state that the main elements of capitalism are "private property in 
means of production and regulation of the productive process by pri- 
vate contract (or management or initiative)." Schumpeter adds to 
this definition of his the factor of "credit creation" (73, p. 167). For 
purposes of sociological analysis it seems advisable — if we want to 
retain the concept of a capitalist society at all — to insist on the union 
of private ownership and factual control of the instruments of pro- 
duction as the distinguishing feature of a capitalist form of society. 
Thus we follow Sering in expecting the "typical capitalist" to be "at 
the same time the legal owner of his factory, the practical manager 
of production, and the supreme commander of his workers" (74, p. 

Both terms, "capitalism" and "industrial society," are categories 
on a high level of generality. In analyzing particular societies they 
are applicable, but not very useful. They are models describing 
merely the most general features of social structures.^ If our aim 
were a "complete" description or analysis of any given society, we 

^ Without here continuing an unconcluded discussion, it may be said for the sake 
of clarity that "models of structure" in this sense differ from Weber's "ideal types" 
in that they do not "idealize" existing conditions in a pure form (and thereby assume 
a goal character, if without explicit evaluation), but indicate merely the common 
factual skeleton of comparable societies without taking into account their cultural 

Changes Since Marx 41 

would have to qualify its classification as a "capitalist" or an "indus- 
trial" society, and add many a reference to particular structural vari- 
ants, cultural traits, and historical traditions of the society in question. 
If, however, we are concerned merely with the analysis of certain gen- 
eral structural elements, or with that of a multitude of societies that 
share such elements, the categories of "capitalist" and "industrial" 
society may be sufficient in themselves. The title of the present study 
should be understood in this sense. Its subject is not primarily any 
particular society, but the conditions of class formation and class con- 
flict in industrial society, i.e., in all societies which fulfill the general 
conditions of this model of structure. 

Having separated the two groups of elements of which most of 
the traditional definitions of capitalism are composed, we are now able 
to contrast capitalist societies with industrial societies. Following our 
definitions, this means of course contrasting a whole (industrial so- 
ciety) with one of its parts (capitalist society). It is not our intention 
here to add another colorful name to the multitude of terms coined 
for postcapitalist societies in recent years. Faced with the choice be- 
tween "socialist society," "managerial society," "bureaucratic society," 
"advanced industrial society," and other such terms, I have no particu- 
lar preference. In any case no decision is needed here, for the purpose 
of a confrontation of capitalist and industrial society is evidently one 
of generalization. In the argument of this section I have deliberately 
not referred to classes and class conflicts. One of the main questions 
which the present investigation is supposed to answer is: Do classes 
and class conflicts belong to that group of phenomena by which only 
the capitalist type of industrial society is characterized, or is their exist- 
ence a consequence of industrial production itself, and are they there- 
fore a lasting feature of industrial societies? This question will ac- 
company us throughout the following analysis of changes in the struc- 
ture of industrial societies since Marx. 


Marx was right in seeking the root of social change in capitalist 
society in the sphere of industrial production, but the direction these 
changes took turned out to be directly contrary to Marx's expectations. 
With respect to capital, he had, in his later years, at least a vision of 
what was going to happen, as his brief and somewhat puzzled analysis 
of joint-stock companies shows. Joint-stock companies were legally 
recognized in Germany, England, France, and the United States in 
the second half of the nineteenth century. Laws often indicate the 

42 Changes Since Marx 

conclusion of social developments, and indeed early forms of joint- 
stock companies can be traced back at least to the commercial com- 
panies and trade societies of the seventeenth century. But it was in 
the nineteenth and early twentieth centuries that this type of enter- 
prise first gained wide recognition and expanded into all branches of 
economic activity. Today, more than two-thirds of all companies in 
advanced industrial societies are joint-stock companies, and their prop- 
erty exceeds four-fifths of the total property in economic enterprises. 
The enterprise owned and run by an individual, or even a family, has 
long ceased to be the dominant pattern of economic organization. 
Moreover, the stock of companies is dispersed fairly widely. Three 
per cent of the adult population of the Federal Republic of Germany, 
and approximately 8 per cent of that of the United States, own one 
or more shares of joint-stock companies. Probably the proportion in 
other countries is somewhere between these extremes.* For purposes 
of the present analysis, we may add to joint-stock companies the co- 
operative enterprises and those owned by the state, which command 
an ever-increasing proportion of the national wealth in contemporary 
societies. All these together, and their growth in the last decade, leave 
little doubt about the significance of this change. 

It is not surprising that sociologists should have shared, from an 
early date, the interest of lawyers and economists in these new and 
rapidly expanding types of organization. There is, moreover, on the 
whole an astonishing degree of consensus among sociologists on the 
implications of joint-stock companies for the structure of industrial 
enterprises, and for the wider structure of society. If one wants to 
distinguish between points of view, one might contrast a rather more 
radical with a somewhat conservative interpretation of this phenome- 
non. Marx was, in this sense, the founder of the radical school j sur- 
prisingly enough, however, most of his later adherents took the more 
conservative view. 

According to the radical view, joint-stock companies involve a 
complete break with earlier capitalist traditions. By separating what 
has come to be called ownership and control, they give rise to a new 
group of managers who are utterly different from their capitalist 
predecessors. Thus for Marx, the joint-stock company involves a 
complete alienation of capital "from the real producers, and its oppo- 
sition as alien property to all individuals really participating in pro- 

* For the data presented or implied in this paragraph, and in this section in gen- 
eral, see, among other sources, Berle and Means (173), Parkinson (180), Rosenstiel 
(183), Schwantag (184), and the Yearbook of the Institut für Demoskopie (229). 

Changes Since Marx 43 

duction, from the manager down to the last day-laborer" (12, III, 
p. 478). In other words, by separating ownership and control, the 
joint-stock company reduces the distance between manager and worker 
while at the same time removing the owners altogether from the 
sphere of production and thereby isolating their function as exploiters 
of others. It is merely a step from this kind of analysis to the thesis 
that, as Renner has it, the "capitalists without function" yield to the 
"functionaries without capital," and that this new ruling group of 
industry bears little resemblance to the old "full capitalists" (71, pp. 
182, 198). Burnham, Geiger, Sering, and others followed Marx (and 
Renner) in this radical interpretation of the social effects of joint- 
stock companies. 

The conservative view, on the other hand, holds that the conse- 
quences of the apparent separation of ownership and control have 
been vastly overrated. It is argued that in fact owners and controllers, 
i.e., stockholders and managers, are a fairly homogeneous group. 
There are often direct connections between them, and where this is 
not the case, their outlook is sufficiently similar to justify insisting on 
the old assumption of a homogeneous class of capitalists opposed to 
an equally homogeneous class of laborers. This view is not often 
heard in the West nowadays, although traces of it are evident in the 
work of C. Wright Mills (62, 63). It may be added that this con- 
servative view is clearly contrary to Marx's own analysis. 

We cannot here exhaust the complex subject of ownership and 
control, but it seems desirable not to leave the subject without consid- 
ering which of these two views seems more plausible and appropriate. 
There can be little doubt that the social structure of joint-stock com- 
panies as well as cooperative and state-owned enterprises differs from 
that of the classical capitalist enterprise, and that therefore a transition 
from the latter to the former is a process of social change. However, 
what type of change are we dealing with in this problem? Is it a 
change involving the transference of certain rights and duties attached 
to social positions from an old to a new group? Or is it a change that 
involves some rearrangement of the positions endowed with rights 
and duties themselves? These questions are not quite as rhetorical as 
they may sound. In fact, I would claim that the separation of owner- 
ship and control involved both a change in the structure of social posi- 
tions and a change in the recruitment of personnel to these positions. 
But it is evident that, in the first place, joint-stock companies differ 
from capitalist enterprises in the structure of their leading positions. 
In the sphere with which we are here concerned, the process of transi- 

44 Changes Since Marx 

tion from capitalist enterprises to joint-stock companies can be de- 
scribed as a process of role differentiation. The roles of owner and 
manager, originally combined in the position of the capitalist, have 
been separated and distributed over two positions, those of stockholder 
and executive.^ 

At the very least, this process of differentiation means that two 
physical entities occupy the positions formerly occupied by one. But 
this is not all. Apart from its manifest effects, the separation of owner- 
ship and control has a number of latent effects of even greater im- 
portance j i.e., it seems clear that the resulting positions, those of stock- 
holder and executive, differ not only with respect to the obvious rights 
and duties of their incumbents, but also in other respects. Generally, 
the "capitalist without function" is indeed, as Marx emphasized, 
alienated from production, i.e., largely removed from the enterprise 
whose stock he owns. He does not participate in the day-to-day life 
of the enterprise, and above all he does not have a defined place in the 
formal hierarchy of authority in the enterprise. The "functionary 
without capital," on the other hand, has this place, although he typi- 
cally has no property in the enterprise which he runs." 

From the point of view of the social structure of industrial enter- 
prises, this means a significant change in the basis of legitimacy of 
entrepreneurial authority. The old-style capitalist exercised authority 
because he owned the instruments of production. The exercise of 
authority was part and parcel of his property rights, as indeed prop- 
erty may always be regarded from one point of view as simply an 
institutionalized form of authority over others. By contrast to this 
legitimation by property, the authority of the manager resembles in 
many ways that of the heads of political institutions. It is true that 
even for the manager property has not ceased to function as a basis of 
authority. The right of the manager to command and expect obedience 
accrues in part from the property rights delegated to him by the share- 
holders, acting either as a group or through an elected board of direc- 

° In a more extensive analysis, it would have to be recognized that this process 
of differentiation involved not only two, hut at least three, roles and positions: the 
third being that of investor or "finance capitalist" (Hilfcrding). Renner would add 
even more: "Three capitalist character masks have stepped into the place of the one 
full capitalist: the producer owning the capital of production in the mask of the entre- 
preneur, the commercial capitalist in the character mask of the businessman, and the 
financial capitalist" (71, p. 175). 

® In 1935, members of management in i 55 of the 200 largest American corpora- 
tions had on the average no more than i . 74 per cent of the ordinary shares of their 
enterprises (cf. 159, p. 135). 

Changes Since Marx 45 

tors. But besides these delegated property rights, the manager, by 
virtue of his more immediate contact with the participants of produc- 
tion, has to seek a second, and often more important, basis of legiti- 
macy for his authority, namely, some kind of consensus among those 
who are bound to obey his commands. Typically, this consensus merely 
takes the form of an absence of dissensus. However, the manager, 
unlike the "full capitalist," can ill afford to exercise his authority in 
direct and deliberate contravention to the wishes and interests of his 
subordinates. The mechanisms by which manual and clerical workers 
who object to a member of top management can make their interests 
felt are complex and largely unregulated.^ But there are such mecha- 
nisms, and managers have ways and means to forestall their being 
brought to bear. In this sense, the "human relations" movement is 
nothing but a symptom of the changing basis of legitimacy of entre- 
preneurial authority once ownership and control are separated.* 

With the differentiation of capitalist roles, the composition of the 
entrepreneurial class — if it is a class — changes too. This is probably 
a gradual development, but one that is far advanced in most of the 
highly industrialized societies today. If we follow Bendix (138, p. 
228) in distinguishing capitalists, heirs, and bureaucrats as three types 
of entrepreneurs, it is evident that three significantly different pat- 
terns of recruitment correspond to these types. The capitalist in this 
sense is a man who owns and manages an enterprise which he has 
founded himself. From having been perhaps a skilled craftsman or 
a shopkeeper at the beginning of his career, he has built up, "from 
scratch," a sizable firm or factory and one that continues to grow in 
scope, size, and production. The heir, by contrast, is born into the 
ownership of an enterprise, and apart from perhaps a few years' ex- 
perience in some of its departments he has known nothing but the 
property he has inherited. Both the capitalist and the heir are owner- 

^ They extend from direct pressure aimed at forcing the manager to resign or 
change his attitudes to indirect means of disturbing the operation of the enterprise 
which may result in the manager's being reprimanded or deposed by the directors, 
who, in this case, act in a sense on behalf of the employees. 

* Bendix has impressively demonstrated that this change was accompanied bv a 
change in "managerial ideologies," i.e., in attempts to justify theoretically the author- 
ity of the entrepreneur (cf. 138). He asserts an ideological change from basing un- 
limited authority on the interests of a ruling class to presuming, on the part of modern 
management, an identity of interests among all participants of production. Bendix's 
argument lends considerable support to my thesis that the separation of ownership 
and control involved a change in the basis of legitimacy of authority. 

46 Changes Since Marx 

For mere managers, however, there are two typical patterns of 
recruitment, and both of them differ radically from those of capital- 
ists and heirs. One of these patterns is the bureaucratic career. In the 
early joint-stock companies in particular, executives were often chosen 
from among the firm's leading employees, both technical and clerical. 
They had worked their way up from the ranks. More recently, a dif- 
ferent pattern has gained increasing importance. Today, a majority 
of top management officials in industrial enterprises have acquired 
their positions on the strength of some specialized education, and of 
university degrees. Lawyers, economists, and engineers often enter 
management almost immediately after they have completed their 
education, andgradually rise to the top positions. There can be little 
doubt that both these patterns of recruitment — but in particular the 
latter — distinguish managerial groups significantly from those of old- 
style owner-managers as well as new-style mere owners. Their social 
background and experience place these groups into different fields of 
reference, and it seems at least likely that the group of professionally 
trained managers "increasingly develops its own functionally deter- 
mined character traits and modes of thought" (Sering, 74, p. 205). 
For this is, in our context, the crucial effect of the separation of owner- 
ship and control in industry: that it produces two sets of roles the 
incumbents of which increasingly move apart in their outlook on and 
attitudes toward society in general and toward the enterprise in par- 
ticular. Their reference groups differ, and different reference groups 
make for different values. Among classical capitalists, the "organiza- 
tion man" is an unthinkable absurdity. Yet the manager is "not 
the individualist but the man who works through others for others" 
(Whyte, 169, p. 21). Never has the imputation of a profit motive 
been further from the real motives of men than it is for modern 
bureaucratic managers,^ Economically, managers are interested in 
such things as rentability, efficiency, and productivity. But all these 
are indissolubly linked with the imponderables of what has been called 
the social "climate of the enterprise" {Betriebsklima). The manager 
shares with the capitalist two important social reference groups: his 
peers and his subordinates. But his attitude toward these differs con- 
siderably from that of the capitalist (as does consequently, the atti- 

® Which is not to say that this "imputation" may not be useful as an assumption 
in economic theory. There is no need for theoretical assumptions to be altogether 
realistic. However, it may be advisable even for economists to try and extend their 
models by including in their assumptions some of the social factors characteristic of 
entrepreneurial roles. 

Changes Since Marx 47 

tude expected from him by his peers) . For him, to be successful means 
to be liked, and to be liked means, in many ways, to be alike. The 
manager is an involuntary ruler, and his attitudes betray his feelings. 

Before concluding this analysis it would perhaps be well to point 
out briefly what it does not mean or imply. Despite many differences, 
there are without doubt considerable similarities in the positions, roles, 
and attitudes of both the capitalist and the manager. Both are entre- 
preneurial roles, and both are therefore subject to certain expectations 
which no social context can remove. Moreover, there are numerous 
personal and social ties between owners and managers in all industrial 
societies. If anything, the unpropertied managers are more active in 
political affairs, both as individuals and through their associations and 
lobbies. Also, while the joint-stock company has conquered the 
sphere of industrial production (i.e., of secondary industries), it is still 
of only minor importance in the tertiary industries of trade and com- 
merce and in the services. Thus, the separation of ownership and 
control is not as fundamental a change as, say, the industrial revolu- 
tion. But it is a change, and one with very definite, if restricted, im- 
plications for class structure and conflict. 

There is little reason to follow Marx and describe the condition 
of separation of ownership and control as a transitional form of his^ 
torical development. It is no more transitional than any other stage 
of history, and it has already proven quite a vital pattern of social and 
economic structure. But I think that we can follow Marx in his radical 
interpretation of this phenomenon. The separation of ownership and 
control has replaced one group by two whose positions, roles, and 
outlooks are far from identical. In taking this view, one does of course 
agree with Marx against himself. For it follows from this that the 
homogeneous capitalist class predicted by Marx has in fact not devel- 
oped. Capital — and thereby capitalism — has dissolved and given way, 
in the economic sphere, to a plurality of partly agreed, partly com- 
peting, and partly simply different groups. The effect of this develop- 
ment on class conflict is threefold: first, the replacement of capitalists 
by managers involves a change in the composition of the groups par- 
ticipating in conflict J second, and as a consequence of this change in 
recruitment and composition, there is a change in the nature of the 
issues that cause conflicts, for the interests of the functionaries without 
capital differ from those of full-blown capitalists, and so therefore 
do the interests of labor vis-a-vis their new oppo'^dnts; and third, 
the decomposition of capital involves a change in tit-e patterns of con- 
flict. One might question whether this new conflict, in which labor is 

48 Changes Since "Marx 

no longer opposed to a homogeneous capitalist class, can still be de- 
scribed as a class conflict at all. In any case, it is different from the 
division of the whole society into two great and homogeneous hostile 
camps with which Marx was concerned. While I would follow the 
radical view of the separation of ownership and control in industry 
to this point, there is one thing to be said in favor of the conservative 
view. Changes in the composition of conflict groups, of the issues, 
and of patterns of conflict do not imply the abolition of conflict or 
even of the specific conflict between management and labor in industry. 
Despite the effects of the decomposition of capital on class structure, 
we have no reason to believe that antagonisms and clashes of interest 
have now been banned from industrial enterprises. 


While Marx had at least a premonition of things to come with 
respect to capital, he remained unaware of developments affecting the 
unity and homogeneity of labor. Yet in this respect, too, the sphere of 
production which loomed so large in Marx's analyses became the start- 
ing point of changes that clearly refute his predictions. The working 
class of today, far from being a homogeneous group of equally un- 
skilled and impoverished people, is in fact a stratum differentiated by 
numerous subtle and not-so-subtle distinctions. Here, too, history has 
dissolved one position, or role, and has substituted for it a plurality 
of roles that are endowed with diverging and often conflicting ex- 

In trying to derive his prediction of the growing homogeneity of 
labor from the assumption that the technical development of industry 
would tend to abolish all differences of skill and qualification, Marx 
was a genuine child of his century. Only the earliest political econo- 
mists had believed that the division of labor in manufacturing would 
make for an "increase of dexterity in every particular workman" 
(Adam Smith, 1 85, p. 7) by allowing him to refine the "skill acquired 
by frequent repetition of the same process" (Babbage, 172, p. 134). 
Already in the following generation, social scientists were quite unani- 
mous in believing that the processes of industrial production "effect 
a substitution of labor comparatively unskilled, for that which is more 
skilled" (Ure, 1 88, p. 30), and that the division of labor had reached 
a phase "in which we have seen the skill of the worker decrease at the 
rate at which industry becomes more perfect" (Proudhon, 181, p. 
153). Marx was only too glad to adopt this view which tallied so 
well with his general theories of class structure: "The interests and 
life situations of the proletariat are more and more equalized, since 

Changes Since Marx 49 

the machinery increasingly obliterates the differences of labor and 
depresses the wage almost everywhere to an equally low level" (14, 
p. 17). "The hierarchy of specialized workmen that characterizes 
manufacture is replaced, in the automatic factory, by a tendency to 
equalize and reduce to one and the same level every kind of work that 
has to be done" ( 1 2, 1, p. 490) . 

Indeed, so far as we can tell from available evidence, there was, 
up to the end of the nineteenth century, a tendency for most industrial 
workers to become unskilled, i.e., to be reduced to the same low level 
of skill. But since then, two new patterns have emerged which are 
closely related on the one hand to technical innovations in production, 
on the other hand to a new philosophy of industrial organization as 
symbolized by the works of F. W. Taylor (167) and H. Fayol (145). 
First, there emerged, around the turn of the century, a new category 
of workers which today is usually described as semiskilled. As early 
as 1905, Max Weber referred to the growing importance of "the 
semiskilled workers trained directly on the job" (32, p. 502). By the 
1930's, the theory had become almost commonplace that "there is a 
tendency for all manual laborers to become semiskilled machine 
minders, and for highly skilled as well as unskilled workers to become 
relatively less important" (Carr-Saunders and Jones, 89, p. 61). The 
semiskilled differ from the unskilled not so much in the technical 
qualifications required from them for their work, as in certain less 
easily defined extrafunctional skills which relate to their capacity to 
accept responsibility, to adapt to difficult conditions, and to perform a 
job intelligently. These extrafunctional skills are acquired not by 
formal training (although many semiskilled workers receive this 
also), but by experience on the jobj yet these "skills of responsibility" 
constitute a clear line of demarcation between those who have them 
and the unskilled who lack both training and experience. Apart from 
the semiskilled, there appeared, more recently, a new and ever-grow- 
ing demand for highly skilled workers of the engineer type in indus- 
try. Carr-Saunders and Jones, in their statement above, still expected 
the simultaneous reduction of unskilled as well as skilled labor. Today 
we know — as Friedmann (146), Geiger (46), Moore (157), and 
others have pointed out — that the second half of this expectation has 
not come true. Increasingly complex machines require increasingly 
qualified designers, builders, maintenance and repair men, and even 
minders, so that Drucker extrapolates only slightly when he says: 
"Within the working class a new shift from unskilled to skilled labor 
has begun — reversing the trend of the last fifty years. The unskilled 
worker is actually an engineering imperfection, as unskilled work, at 

50 Changes Since Marx 

least in theory, can always be done better, faster and cheaper by ma- 
chines" (144, pp. 42-43). 

Because of changing classifications, it is a little difficult to docu- 
ment this development statistically. As for the unskilled, a slight 
decrease in their proportion can be shown for England where, in 1 9 5 1 , 
they amounted to 12.5 per cent of the occupied male population, as 
against 16.5 per cent in 1 93 1 . In the United States, an even sharper 
decrease has been noted, from 36 per cent of the labor force in 1 9 10 to 
just over 28 per cent in 1930 and, further, to less than 20 per cent in 
1950 (see Caplow, 141, p. 299). But statistics are here neither very 
reliable nor even indispensable evidence. Analysis of industrial condi- 
tions suggests quite clearly that within the labor force of advanced 
industry we have to distinguish at least three skill groups: a growing 
stratum of highly skilled workmen who increasingly merge with both 
engineers and white-collar employees, a relatively stable stratum of 
semiskilled workers with a high degree of diffuse as well as specific 
industrial experience, and a dwindling stratum of totally unskilled 
laborers who are characteristically either newcomers to industry (be- 
ginners, former agricultural laborers, immigrants) or semi-unem- 
ployables. It appears, furthermore, that these three groups differ not 
only in their level of skill, but also in other attributes and determi- 
nants of social status. The semiskilled almost invariably earn a higher 
wage than the unskilled, whereas the skilled are often salaried and 
thereby participate in white-collar status. The hierarchy of skill cor- 
responds exactly to the hierarchy of responsibility and delegated au- 
thority within the working class. From numerous studies it would 
seem beyond doubt that it also correlates with the hierarchy of pres- 
tige, at the top of which we find the skilled man whose prolonged 
training, salary, and security convey special status, and at the bottom 
of which stands the unskilled man who is, according to a recent Ger- 
man investigation into workers' opinions, merely "working" without 
having an "occupation" proper (see Kluth, 1 50, p. 67) . Here as else- 
where Marx was evidently mistaken. "Everywhere, the working 
class differentiates itself more and more, on the one hand into occu- 
pational groups, on the other hand into three large categories with 
different, if not contradictory, interests: the skilled craftsmen, the 
unskilled laborers, and the semiskilled specialist workers" (Philip, 
161, p. 2).^° 

In trying to assess the consequences of this development, it is well 

*° In argument and evidence, the preceding account is based on two more elabo- 
rate studies of mine, one of unskilled labor (143), and one of skill and social stratifi- 
cation (142). For further references as well as data I must refer to these studies. 

Changes Since Marx 5 1 

to remember that, for Marx, the increasing uniformity of the working 
class was an indispensable condition of that intensification of the class 
struggle which was to lead, eventually, to its climax in a revolution. 
The underlying argument of what for Marx became a prediction ap- 
pears quite plausible. For there to be a revolution, the conflicts within 
a society have to become extremely intense. For conflicts to be intense, 
one would indeed expect its participants to be highly unified and ho- 
mogeneous groups. But neither capital nor labor have developed 
along these lines. Capital has dissolved into at least two, in many ways 
distinct, elements, and so has labor. The proletarian, the impoverished 
slave of industry who is indistinguishable from his peers in terms of 
his work, his skill, his wage, and his prestige, has left the scene. What 
is more, it appears that by now he has been followed by his less de- 
praved, but equally alienated successor, the worker. In modern indus- 
try, "the worker" has become precisely the kind of abstraction which 
Marx quite justly resented so much. In his place, we find a plurality 
of status and skill groups whose interests often diverge. Demands 
of the skilled for security may injure the semiskilled; wage claims 
of the semiskilled may raise objections by the skilled; and any interest 
on the part of the unskilled is bound to set their more highly skilled 
fellow workmen worrying about diflFerentials. 

Again, as in the case of capital, it does not follow from the decom- 
position of labor that there is no bond left that unites most workers 
— at least for specific goals; nor does it follow that industrial conflict 
has lost its edge. But here, too, a change of the issues and, above all, 
of the patterns of conflict is indicated. As with the capitalist class, it 
has become doubtful whether speaking of the working class still makes 
much sense. Probably Marx would have agreed that class "is a force 
that unites into groups people who differ from one another, by over- 
riding the differences between them" (Marshall, 57, p. 114), but he 
certainly did not expect the differences to be so great, and the uniting 
force so precarious as it has turned out to be in the case both of capital 
and of labor. 

THE "new middle CLASs" 

Along with the decomposition of both capital and labor a new 
stratum emerged within, as well as outside, the industry of modern 
societies, which was, so to speak, born decomposed. Since Lederer and 
Marschak first published their essay on this group, and coined for it 
the name "new middle class" {neuer Mittelstand) ^ so much has been 
written by sociologists about the origin, development, position, and 
function of white-collar or black-coated employees that whatever one 

52 Changes Since Marx 

says is bound to be repetitive. However, only one conclusion is borne 
out quite clearly by all these studies of salaried employees in industry, 
trade, commerce, and public administration: that there is no word in 
any modern language to describe this group that is no group, class that 
is no class, and stratum that is no stratum. To be sure, there have been 
attempts to describe it. In fact, we are here in the comparatively 
fortunate position of having to decide between two or, perhaps, three 
conflicting theories. But none of these attempts has been free of in- 
numerable qualifications to the effect that it is impossible to generalize. 
Although the following brief discussion will not distinguish itself in 
this respect, it could not be avoided in an account of social changes of 
the past century that have a bearing on the problem of class. 

By the time Marx died, about one out of every twenty members 
of the labor force was in what might roughly be described a clerical 
occupation j today, it is one out of every five and, in the tertiary indus- 
tries, one out of every three. More accurate figures of size and growth 
of the "new middle class" could be given,^^ but even these are sur- 
prisingly precise in view of the fact that it is virtually impossible to 
delimit the "group" which they count. For, technically, the "occupa- 
tional salad" (Mills) of salaried employees includes post-office clerks 
as well as senior executives, shop supervisors as well as hospital doc- 
tors, typists as well as prime ministers. Presumably, a "middle class" 
is located somewhere between at least two other classes, one above it 
and one below it. Yet the "new middle class" has stubbornly resisted 
all attempts to define its upper and lower limits. In fact, it is obvious 
that the questions where salaried employees begin to be members of 
an upper stratum or ruling class and where they "really" still belong 
in the working class cannot in general be answered. Our questions 
will have to be rather more specific. 

If one is, as we are, concerned not with patterns of social stratifi- 
cation but with lines of conflict, then one thing is certain : however we 
may choose to delimit the aggregate of salaried employees, they are 
not a "middle class," because from the point of view of a theory of 
conflict there can be no such entity as a middle class. Evident as it is, 
this statement is bound to be misunderstood — but, then, much of this 
study is an attempt to elucidate it. It is true that in terms of prestige 
and income many salaried employees occupy a position somewhere 

^^ For data used in the following section, see, apart from the early work by 
Lederer and Marschak (133), above all the studies of Lockwood (135), Geiger (91), 
Croner (129), Lewis and Maude (134), and C. Wright Mills (137). Some in- 
teresting figures are given by Bendix (138, p. 214). 

Changes Since Marx 53 

between the very wealthy and the very poor, somewhere in the middle 
of the scale of social stratification. But in a situation of conflict, 
whether defined in a Marxian way or in some other way, this kind of 
intermediate position just does not exist, or, at least, exists only as a 
negative position of nonparticipation. This point might be illuminated 
by a slightly misleading example: an election in which there is a choice 
between two parties j while it is possible to abstain, only those who 
make up their minds one way or the other participate actively in the 
contest. Similarly, our problem here is to determine in which way the 
so-called new middle class has made up its mind or is likely to make 
up its mind. And the answer we shall give corresponds — to remain 
within the metaphor for a moment — to the findings of Bonham in 
England (127) and Von der Heydte in Germany (196), according 
to which two-thirds of the "new middle class" tend to vote for con- 
servative, and one-third for radical parties. 

I have claimed above that there are two or, perhaps, three com- 
peting theories about the position of the "new middle class." From 
the point of view of our problem, these are soon reduced to what at 
best amounts to two-and-a-half theories. For the third theory I had 
in mind is in fact little more than a description, and an inconclusive 
one at that. It is embodied in Crozier's "working hypothesis" of an 
empirical investigation conducted in France: "The situation of the 
salaried employee is one that makes possible an identification with the 
world of the ruling class and promises considerable rewards if this 
succeeds. But at the same time it is a working-class situation and there- 
fore suffers from most of those limitations to which all other workers 
are subjected — limited income as well as lack of autonomy and a 
position of subordination" ( 1 30, pp. 3 1 i-i 2) . Statements of this kind 
are as frequent as they are useless for purposes of conflict analysis. 
We can therefore dismiss at the outset any theory that confines itself 
to statements with clauses like "partly this . . . partly that" or "on 
the one hand this ... on the other hand that." 

There are two theories which do not suffer from this indecision, 
and they are directly contradictory. According to the first of these, 
the "new middle class" constitutes in fact an extension of the old, capi- 
talist or bourgeois, ruling class, and is in this sense part of the ruling 
class. Croner — who, apart from Renner (71), Bendix (138, chap, 
iv), and others, recently espoused this theory — argues that "the expla- 
nation of the special social position of salaried employees can be found 
in the fact that their work tasks have once been entrepreneurial tasks" 
(129, p. T^G). This statement is meant by Croner both in a historical 

54 Changes Since Marx 

and in a structural sense. Historically, most clerical occupations were 
differentiated out of the leading positions in industry, commerce, and 
the state. Structurally they are, according to this view, characterized 
by the exercise of delegated authority — delegated, that is, from the 
real seat of authority in social organizations, from, in other words, 
their leading positions. In contrast to this view, Geiger, C. Wright 
Mills, and others claim that the "new middle class" is, if not exactly 
an extension of the proletariat, at any rate closer to the working class 
than to the ruling class, whether capitalist or managerial. "Objec- 
tively, . . . the structural position of the white-collar mass is becom- 
ing more and more similar to that of the wage-workers. Both are, of 
course, propertyless, and their incomes draw closer and closer to- 
gether. All the factors of their status position, which have enabled 
white-collar workers to set themselves apart from wage-workers, are 
now subject to definite decline" (137, p. 297). Mills does not say so, 
but he would probably have no quarrel with Geiger's conclusion that 
"from the point of view of class structure in Marx's sense the salaried 
employee is undoubtedly closer to the worker than to any other figure 
of modern society" (46, p. 167). 

The two views are clearly in conflict, and it seems desirable to come 
to a decision as to their relative merits.^" Fortunate as it is, from a 
methodological point of view, to have to decide between two conflict- 
ing theories, our situation here does not, upon closer inspection, turn 
out to be quite so simple. In fact. Mills may well be right when he 
suspects that because of the vastly different "definitions" of the "new 
middle class" the two theories not only may peacefully co-exist but 
even both be correct (137, pp. 291 ff.). Clearly the theory that sal- 
aried employees have delegated authority and are therefore part of 
the ruling class cannot have meant the office boy, the salesgirl, or even 
the skilled worker who has been granted the status symbol of a salary; 
equally clearly, the theory that salaried employees resemble the work- 
ing class does not apply to senior executives, higher civil servants, and 

^^ C. Wright Mills, in his very balanced account of views about the "new middle 
class" (137, chap. 13), enumerates four competing theories (pp. 290 ff.). To the two 
mentioned he adds the theories that (l) the middle class is destined to be the ruling 
class of the future, and (2) the growth of the "new middle class" operates as a force 
to stabilize the old, and eventually to abolish all class conflicts. The latter view will 
be discussed at a later stage, since it presupposes some consideration of the significance 
of social conflict in general. As to the former view, it simply betrays an unpardonable 
confusion of terms on the part of the authors Mills refers to. It seems obvious that so 
long as the middle class is a middle class there must be a class above it, and once it is 
the ruling class it is no longer the middle class. 

Changes Since Marx SS 

professional people. However, there is more than a question of defi- 
nition involved in this difficulty. 

Instead of asking which of two apparently conflicting theories ap- 
plies to the "new middle class," we can, so to speak, reverse our ques- 
tion and ask whether there is any criterion that would allow us to dis- 
tinguish between those sectors of the "new middle class" to which one 
theory applies and those to which the other theory applies. I think 
that there is such a criterion, and that its application provides at least a 
preliminary solution to our wider problem of the effects of the growth 
of a "new middle class" on class structure and class conflict. It seems 
to me that a fairly clear as well as significant line can be drawn between 
salaried employees who occupy positions that are part of a bureau- 
cratic hierarchy and salaried employees in positions that are not. The 
occupations of the post-office clerk, the accountant, and, of course, 
the senior executive are rungs on a ladder of bureaucratic positions; 
those of the salesgirl and the craftsman are not. There may be barriers 
in bureaucratic hierarchies which are insurmountable for people who 
started in low positions; salaried employees outside such hierarchies 
may earn more than those within, and they may also change occupa- 
tions and enter upon a bureaucratic career; but these and similar facts 
are irrelevant to the distinction between bureaucrats and white-collar 
workers proposed here. Despite these facts I suggest that the ruling- 
class theory applies without exception to the social position of bureau- 
crats, and the working-class theory equally generally to the social posi- 
tion of white-collar workers. 

There is, in other words, one section of the "new middle class" 
the condition of which, from the point of view of class conflict, closely 
resembles that of industrial workers. This section includes many of 
the salaried employees in the tertiary industries, in shops and restau- 
rants, in cinemas, and in commercial firms, as well as those highly 
skilled workers and foremen who have acquired salaried status. It is 
hard to estimate, from available evidence, the numerical size of this 
group, but it probably does not at present exceed one-third of the 
whole "new middle class" — although it may do so in the future, since 
the introduction of office machinery tends to reduce the number of 
bureaucrats while increasing the demand for salaried office techni- 
cians.^^ Although some white-collar workers earn rather more than 

^^ It is still too early to make any definite statements about this important devel- 
opment, which Bahrdt described as the "industrialization of bureaucracy" (cf. 124) ; 
but the automation of office work is sure to have consequences for the class structure of 
contemporary societies. 

ß6 Changes Since Marx 

industrial workers, and most of them enjoy a somewhat higher pres- 
tige, their class situation appears sufficiently similar to that of workers 
to expect them to act alike. In general, it is among white-collar work- 
ers that one would expect trade unions as well as radical political par- 
ties to be successful. 

The bureaucrats, on the other hand, share, if often in a minor way, 
the requisites of a ruling class. Although many of them earn less than 
white-collar and even industrial workers, they participate in the exer- 
cise of authority and thereby occupy a position vis-a-vis rather than 
inside the working class. The otherwise surprising fact that many 
salaried employees identify themselves with the interests, attitudes, 
and styles of life of the higher-ups can be accounted for in these terms. 
For the bureaucrats, the supreme social reality is their career that pro- 
vides, at least in theory, a direct link between every one of them and 
the top positions which may be described as the ultimate seat of au- 
thority. It would be false to say that the bureaucrats are a ruling class, 
but in any case they are part of it, and one would therefore expect 
them to act accordingly in industrial, social, and political conflicts." 

The decomposition of labor and capital has been the result of social 
developments that have occurred since Marx, but the "new middle 
class" was born decomposed. It neither has been nor is it ever likely 
to be a class in any sense of this term. But while there is no "new 
middle class," there are, of course, white-collar workers and bureau- 
crats, and the growth of these groups is one of the striking features 
of historical development in the past century. What is their effect on 
class structure and class conflict, if it is not that of adding a new class 
to the older ones Marx described? It follows from our analysis that 
the emergence of salaried employees means in the first place an exten- 
sion of the older classes of bourgeoisie and proletariat. The bureau- 
crats add to the bourgeoisie, as the white-collar workers add to the 
proletariat. Both classes have become, by these extensions, even more 
complex and heterogeneous than their decomposition has made them 
in any case. By gaining new elements, their unity has become a highly 
doubtful and precarious feature. White-collar workers, like indus- 
trial workers, have neither property nor authority, yet they display 
many social characteristics that are quite unlike those of the old work- 

^* Evidently, rather more will have to be said to make a convincing case for what 
I have sketched very briefly and somewhat dogmatically at this point. In so far as the 
structural position of bureaucrats has a bearing on the problem of class conflict in the 
present and future, these remarks will in fact be extended and supplemented below 
in Chapter VIII. 

Changes Since Marx 57 

ing class. Similarly, bureaucrats differ from the older ruling class 
despite their share in the exercise of authority. Even more than the 
decomposition of capital and labor, these facts make it highly doubtful 
whether the concept of class is still applicable to the conflict groups 
of post-capitalist societies. In any case, the participants, issues, and 
patterns of conflict have changed, and the pleasing simplicity of 
Marx's view of society has become a nonsensical construction. If ever 
there have been two large, homogeneous, polarized, and identically 
situated social classes, these have certainly ceased to exist today, so 
that an unmodified Marxian theory is bound to fail in explaining the 
structure and conflicts of advanced industrial societies. 


The decomposition of capital and labor as well as their extension 
by sections of the "new middle class" are phenomena which have an 
obvious and direct bearing on class structure. But they are neither the 
only changes that have occurred since Marx nor, perhaps, the most 
significant ones from the point of view of class. Apart from such po- 
litical and economic forces as totalitarianism and socialism, it was in 
particular the institutionalization of the two great social forces of mo- 
bility and equality that has steered class structures and conflicts in 
directions unforeseen by Marx. Marx was not, in fact, unaware of the 
importance of these forces. In explaining the absence of stable classes 
in the United States in terms of what he called the "exchange between 
classes" (cf. 8, p. 18), he anticipated the cardinal thesis of Sombart's 
brilliant essay. Why is there no Socialism in the United States? (204) . 
But for Marx, mobility was a symptom of short-lived transitional 
periods of history, i.e., of either the emergence or the impending 
breakdown of a society. Today, we would tend to take the opposite 
view. Social mobility has become one of the crucial elements of the 
structure of industrial societies, and one would be tempted to predict 
its "breakdown" if the process of mobility were ever seriously im- 
peded. Marx believed that the strength of a ruling class documents 
itself in its ability to absorb the ablest elements of other classes. In 
a manner of speaking, this is permanently the case in advanced indus- 
trial societies, yet we should hesitate to infer from a steady increase 
in the upward mobility of the talented that the present ruling class 
is particularly strong or homogeneous. 

Social mobility reprcsenco one of the most studied and, at the same 
time, least understood areas of sociological inquiry. Today, we know 
a great deal about social mobility in various countries, and yet we do 

5 8 Changes Since Marx 

not really know what we know. Not only do we not have any satis- 
factory answer to the question about the causes and consequences of 
social mobility that was recently put by Lipset and Zetterberg ( 1 1 6, 
p. 158), but we cannot even be sure about the so-called facts of the 
case. The evidence we have is most conclusive with respect to mobility 
between generations, although even here generalizations rest on extra- 
polation as much as on interpretation. It appears that in countries like 
the United States, Great Britain, and Germany the rate of intergen- 
eration mobility is generally fairly high. Only in the highest and, in 
some countries, in the lowest ranges of the occupational scale do we 
still find a considerable amount of self-recruitment. Moreover, the 
rate of mobility seems to correspond roughly to the degree of indus- 
trialization in a country. It is higher in Britain than in France, higher 
in the United States than in Italy. This correlation between industrial 
development and social mobility seems to hold also in the historical 
dimension. For Britain and Germany, investigations suggest a con- 
siderable increase in mobility rates over the last three generations.^^ 
However, even if these generalizations are taken as suggestions 
rather than conclusions, they have a high degree of verisimilitude. 
For with respect to intergeneration mobility we have at our disposal 
another kind of evidence which, although not quantitative, is quite 
conclusive. When Marx wrote his books, he assumed that the position 
an individual occupies in society is determined by his family origin and 
the position of his parents. The sons of workers have no other choice 
but to become workers themselves, and the sons of capitalists stay in 

■^^ The first of these generalizations is really little more than a guess based on the 
interpretation of mobility studies by Glass and others in Britain (107), Bolte in Ger- 
many (102, 103), Rogoff in the United States (121), the Japan Sociological Society 
(115), and the data included in the article by Lipset and Zetterberg (116). Some 
comparative data have been brought together in this article also, as well as in the volume 
edited by Glass (ill, p. 263). Historical studies of mobility suffer from the fact that 
they have to rely on people's memories; some tentative findings have been presented 
by Mukherjee (119, p. 284) and Bolte (104, p. 186). Despite my qualifications, 
this evidence is not, after all, unimpressive. However, all the studies mentioned do 
not really stand up to a thorough methodological inspection. This is almost obvious 
in comparative analyses based on vastly different occupational classifications which no- 
body so far has taicen the trouble to reclassify. But it is also true for studies in one 
country. They have usually employed the index of association for measuring mobility 
rates, and, as Professor John W. Tukey has pointed out to me, this index is neither 
formally sound nor empirically useful. Formally, the index of association would have 
to be weighted by the size of status categories to be of any use at all. Empirically, it 
fails to describe the most important aspect of social mobility: the existence of barriers 
between strata. These and other objections to existing studies make great caution 
imperative in the use of their findings. 

Changes Since Marx $9 

the class of their fathers. At the time, this assumption was probably 
not far from the truth. But since then a new pattern of role allocation 
has become institutionalized in industrial societies. Today, the alloca- 
tion of social positions is increasingly the task of the educational sys- 
tem. Even a hundred years ago, "the attendance of a certain type of 
school meant a confirmation of a certain social status or rank, and not 
its acquisition" (Schelsky, I22, p. 3). Today, the school has become 
the "first and thereby decisive point of social placement with respect 
to future social security, social rank and the extent of future consump- 
tion chances" (122, p. 6). In post-capitalist society, it is "the process 
of socialization itself, especially as found in the educational system, 
that is serving as the proving ground for ability and hence the selective 
agency for placing people in different statuses according to their ca- 
pacities" (Davis, 208, p. 219). To be sure, there still are numerous 
obstacles and barriers in the way of complete equality of educational 
opportunity, but it is the stubborn tendency of modern societies to in- 
stitutionalize intergeneration mobility by making a person's social po- 
sition dependent on his educational achievement. Where this is the 
case, no social stratum, group, or class can remain completely stable for 
more than one generation. Social mobility, which, for Marx, was the 
exception that confirmed the rule of class closure, is built into the 
structure of post-capitalist society and has therefore become a factor 
to be reckoned with in all analyses of conflict and change. 

There are forms of mobility other than that between generations, 
but about these we know even less. Thomas' study of intrageneration 
mobility in Britain (123) suggests a truly extraordinary degree of 
exchange between occupational groups. According to this study, there 
is not in contemporary Britain a single status category the majority of 
the members of which have never been in higher or lower strata (123, 
p. 30) . But while it seems probable that there is a considerable amount 
of movement between occupations in various spheres of work, find- 
ings like those of Thomas's would require a more thorough analysis 
than the data published so far permit. There are sure to be even 
higher barriers for intrageneration mobility than for mobility between 
generations, so that all we can infer from what patchy data we have 
is what anybody living in a modern society can observe for himself: 
in post-capitalist societies there is a great deal of movement, upwards 
and downwards as well as on one social level, between generations as 
well as within them, so that the individual who stays at his place of 
birth and in the occupation of his father throughout his life has be- 
come a rare exception. 

When Marx dealt with the "exchange between classes" in the 

6o Changes Since Marx 

United States, he assumed that a high degree of exchange would be 
detrimental to the formation of powerful classes and therefore incon- 
ducive to the fomentation of violent conflicts. This assumption is 
plausible. A class composed of individuals whose social position is not 
an inherited and inescapable fate, but merely one of a plurality of so- 
cial roles, is not likely to be as powerful a historical force as the closed 
class Marx had in mind. Where mobility within and between genera- 
tions is a regular occurrence, and therefore a legitimate expectation 
of many people, conflict groups are not likely to have either the per- 
manence or the dead seriousness of caste-like classes composed of 
hopelessly alienated men. And as the instability of classes grows, the 
intensity of class conflict is bound to diminish. Instead of advancing 
their claims as members of homogeneous groups, people are more 
likely to compete with each other as individuals for a place in the sun. 
Where such competition is not possible, or not successful, group con- 
flicts assume a somewhat milder and looser character than class strug- 
gles of a Marxian type. Again, the question arises whether such con- 
flict groups of mobile individuals can still be described as classes. In 
any case, the institutionalization of social mobility through both the 
educational and the occupational systems contradicts quite clearly the 
prediction of a continuous increase in the intensity of class conflicts. 
In a study of the effects of social mobility on group relations, 
Janowitz arrived at two interesting conclusions: "One, social mobility 
generally has been found to have disruptive implications for the struc- 
ture of primary group relations and on related social psychological 
states, and thereby to carry socially maladjustive consequences. . . . 
Second [with respect to the consequences of social mobility for second- 
ary group structures], markedly diflFerent order of inferences can be 
made. Upward social mobility, especially in the middle class, tends 
to orient and incorporate mobile groups into many types of secondary 
structures with relative eff^ectiveness. On the other hand, . . . down- 
ward mobility does not produce effective involvement in secondary 
group structures in pursuit of self-interest" ( i I4,p. 193). This find- 
ing (which evidently applies to intergeneration mobility only) is a 
welcome reminder of the fact that although mobility diminishes the 
coherence of groups as well as the intensity of class conflict, it does 
not eliminate either. While civil wars and revolutions may be unlikely 
in a highly mobile society, there is no a priori reason to believe that 
conflicts of interests will not find their expression in other ways. 
Marx's theory of class fails to account for such other types of conflict, 
and it will be our task in this study to find an approach that accounts 

Changes Since Marx 6 1 

for group conflicts in mobile as well as in relatively immobile indus- 
trial societies. 


In the preceding sections two of the three predictions that Marx 
made about the future development of classes in capitalist society- 
have been discussed in the light of the social history of the last decades. 
We have seen that neither of them has come true. Contrary to Marx's 
expectations, the increasing differentiation as well as homogeneity of 
classes was checked by the decomposition of labor and capital, the 
emergence of white-collar workers and bureaucrats, and the institu- 
tionalization of social mobility. But none of Marx's hopes — for such 
they were — has been refuted more dramatically in social development 
than his prediction that the class situations of bourgeoisie and pro- 
letariat would tend toward extremes of wealth and poverty, posses- 
sion and deprivation. Here, too, Marx had a simple theory. He 
believed in a direct and unfailing correlation between the extremity 
of class situations and the intensity of class conflict. It is quite pos- 
sible that this theory contains an element of truth, but if it does, 
then the remarkable spread of social equality in the past century 
has rendered class struggles and revolutionary changes utterly im- 

T. H. Marshall has shown that much of modern social history 
can be understood in terms of what he calls the "war" between "citi- 
zenship rights" (which, by definition, are equal rights) and the "capi- 
talist class system" (cf. 57). In successive periods, three types of 
citizenship rights have been adopted by most industrial societies, and 
they have increasingly affected the processes of class differentiation 
and class conflict. The first of these rights, that of the generalization 
of legal equality, was still quite compatible with class conflict and even 
with class war. Indeed, Marx used his most mocking and cynical style 
when he referred to legal equality in capitalist society: "Liberty! 
For buyers and sellers of a commodity, e.g., labor power, are deter- 
mined by their free will alone. They enter into contracts as free, 
legally equal persons. . . . Equality! For they are related merely 
as owners of commodities, and exchange equivalent for equivalent. 
Property! For everybody controls merely what is his" (12, I, p. 
1 84). But Marx overlooked what Tocqueville (whose work he prob- 
ably knew) had observed before him, namely, that equality is a highly 
dynamic force, and that men, once they are equal in some respects, 
"must come in the end to be equal upon all" (232, p. $$)- 

62 Changes Since Marx 

A considerable step towards complete equality was taken when 
citizenship rights were, in the nineteenth century, extended to the 
political sphere. Universal suffrage and the right to form political 
parties and associations involved the removal of political conflicts 
from the factory floor and the street to negotiating bodies and parlia- 
ments. On a different level, it opened up the possibility for Marx's 
followers to convert their master's theories into political realities — 
but it is as well that they did not fail as miserably in this process as 
Marx did himself. By virtue of freedom of association and political 
equality, the early trade-union movement as well as socialist parties 
grown out of it achieved considerable success in improving the lot of 
the working class, although this progress was still restricted by many 
an obstacle. "Civil rights gave legal powers whose use was drastically 
curtailed by class prejudice and lack of economic opportunity. Politi- 
cal rights gave potential power whose exercise demanded experience, 
organization, and a change of ideas as to the proper functions of gov- 
ernment" (57, p. 46). Only when, in our own century, legal and 
political citizenship rights were supplemented by certain social rights, 
did the process of equalization of status really reach a point where the 
differences and antagonisms of class are affected. 

The social rights of citizenship which are widely recognized in 
contemporary societies include old-age pensions, unemployment bene- 
fits, public health insurance, and legal aid, as well as a minimum wage 
and, indeed, a minimum standard of living. "Equal participation in 
the material and intellectual comforts of civilization ... is the un- 
disputed basic material right of our social constitution" (Schelsky, 
122, p. 5). Where established rights guarantee this kind of equality 
for every citizen, conflicts and differences of class are, at the very least, 
no longer based on inequalities of status in a strict sense of this term. 
From the point of view of legal privileges and deprivations, every 
citizen of advanced industrial societies has an equal status, and what 
social differences there arc arise on the undisputed basis of this funda- 
mental equality. The "absolute" privilege of the bourgeoisie, and the 
equally "absolute" alienation of the proletariat which Marx with a 
characteristic Hegelian figure of thought predicted, has not only not 
come true, but, by institutionalizing certain citizenship rights, post- 
capitalist society has developed a type of social structure that excludes 
both "absolute" and many milder forms of privilege and deprivation. 
If equality before the law was for most people in the early phases of 
capitalist society but a cynical fiction, the extended citizenship rights 

Changes Since Marx 63 

of post-capitalist society represent a reality that forcefully counteracts 
all remaining forms of social inequality and differentiation. 

This is a fortiori the case, since along with the spread of citizen- 
ship rights, the social situation of people became increasingly similar. 
The completeness of this leveling tendency can be, and has been, exag- 
gerated. There are of course even today considerable differences in 
income, prestige, spending habits, and styles of life. But as a tend- 
ency the process of leveling social differences cannot be denied. By 
the simultaneous rise of the real wages of workers and the taxation 
of top earnings, a redistribution of incomes has taken place — a redis- 
tribution that some believe today has gone so far as to remove every 
incentive for work requiring special training or skill. Many of the 
technical comforts and status symbols of modern life are increas- 
ingly available to everybody. The mass-produced commodities of the 
"culture industry" (Horkheimer and Adorno, 227) unite distant peo- 
ple and areas in nearly identical leisure-time activities. Schelsky gives 
voice to the impression of many when he summarizes this develop- 
ment as a process of "social leveling with predominantly petty-bour- 
geois or middle-class patterns of behavior and ideals" (231, p. 349).^" 

Social stratification and class structure are two distinct aspects of 
social organization, but they both refer to inequalities in the social life 
of individuals. If, therefore, the legal and social status of people 
undergoes a process of leveling which apparently tends towards com- 
plete equality of status, the concepts of social stratification and class 
structure tend to lose their meaning. In so far as social stratification 
is concerned, there is evidence and argument to contest this inference. 
For one thing there is some doubt as to whether one can really extra- 
polate from developments of the past century and infer a further 
leveling of socioeconomic status. For another thing, it seems far more 
likely from the point of view of a functioning social structure that 
there is a certain minimum of inequalities which wilJ not be touched 
by egalitarian trends under any condition. ^^ With respect to class 
structure, the answer is not as simple. There can be little doubt that 

^* There is a host of literature on problems of equality and social class. Apart 
from the works of Marshall and Schelsky (which are discussed more extensively in 
Chapter III, below), see especially Tawney's classic (187), and Bottomore's recent 
discussion of the problem (37). 

■^'^ The argument of Davis and Moore (90) to this effect is not, in my opinion, 
wholly convincing, but many of the considerations relevant to the problem of the func- 
tion of inequality have come out in their argument and the controversy following it. 

64 Changes Since Marx 

the equalization of status resulting from social developments of the 
past century has contributed greatly to changing the issues and dimin- 
ishing the intensity of class conflict. By way of extrapolation — fairly 
wild extrapolation, I may say — some authors have visualized a state 
in which there are no classes and no class conflicts, because there is 
simply nothing to quarrel about. I do not think that such a state is 
ever likely to occur. But in order to substantiate this opinion, it is 
necessary to explore the structural limits of equality, i.e., to find the 
points at which even the most fanatic egalitarian comes up against 
insurmountable realities of social structure. One of these is surely 
the variety of human desires, ideas, and interests, the elimination of 
which is neither desirable nor likely. But while this is important, it 
is not as such an element of social structure. I shall suggest in this 
study that the fundamental inequality of social structure, and the 
lasting determinant of social conflict, is the inequality of power and 
authority which inevitably accompanies social organization. But this 
is an anticipation about which much more will have to be said. In so 
far as the theory and practice of equality in post-capitalist societies 
are concerned, it seems certain that they have changed the issues and 
patterns of class conflict, and possible that they have rendered the 
concept of class inapplicable, but they have not removed all significant 
inequalities, and they have not, therefore, eliminated the causes of 
social conflict. 


A historian might argue that all the tendencies of change here 
described as changes in the structure of industrial societies since Marx 
had in fact begun before and in some cases long before Marx died in 
1883. Joint-stock companies existed even before the industrial revo- 
lution j there have always been workers of varying degrees of skill j 
bureaucrats and white-collar workers are not an invention of capital- 
ism, nor is social mobility 5 and the spread of social equality began at 
least with the French Revolution, if not earlier. The historian is of 
course right, but so is, strangely enough, the sociologist (and while the 
latter usually admits the former, the historian is rather less likely to 
admit the latter). The sociologist is generally interested not so much 
in the origin of social phenomena as in their spread and rise to wider 
significance. There is. however, one line of social development in 
industrial societies which has both originated and spread since about 
the time of Marx's death, and which is directly relevant to our prob- 
lem. Geiger, who has described this change as the "institutionalization 

Changes Since Marx 6$ 

of class conflict," says: "The tension between capital and labor is 
recognized as a principle of the structure of the labor market and has 
become a legal institution of society. . . . The methods, weapons, 
and techniques of the class struggle are recognized — and are thereby 
brought under control. The struggle evolves according to certain 
rules of the game. Thereby the class struggle has lost its worst sting, 
it is converted into a legitimate tension between power factors which 
balance each other. Capital and labor struggle with each other, con- 
clude compromises, negotiate solutions, and thereby determine wage 
levels, hours of work, and other conditions of work" (46, p. 1 84). 

Marx displayed a certain sociological naivete when he expressed 
his belief that capitalist society would be entirely unable to cope with 
the class conflict generated by its structure. In fact, every society is 
capable of coping with whatever new phenomena arise in it, if only 
by the simple yet effective inertia which can be described, a little pre- 
tentiously, as the process of institutionalization. In the case of class 
conflict, institutionalization assumed a number of successive and com- 
plementary forms. It began with the painful process of recognition 
of the contending parties as legitimate interest groups. Within indus- 
try, a "secondary system of industrial citizenship" (Marshall, 57, 
p. 68) enabled both workers and entrepreneurs to associate and defend 
their interests collectively. Outside industry, the primary system of 
political citizenship had the same effect. And while, in the stage of 
organization, conflict may develop a greater visible intensity, organi- 
zation has at least two side effects which operate in the opposite direc- 
tion. Organization presupposes the legitimacy of conflict groups, and 
it thereby removes the permanent and incalculable threat of guerrilla 
warfare. At the same time, it makes systematic regulation of conflicts 
possible. Organization is institutionalization, and whereas its mani- 
fest function is usually an increasingly articulate and outspoken de- 
fense of interests, it invariably has the latent function also of inaugu- 
rating routines of conflict which contribute to reducing the violence 
of clashes of interest. 

These are generalizations derived from the experience of class 
conflict in capitalist and post-capitalist societies. Here, the organi- 
zation of capital and labor, bourgeoisie and proletariat, was soon fol- 
lowed by several further patterns of conflict regulation. On the one 
hand, the contending parties in industry and politics agreed on certain 
rules of the game and created institutions which provided a frame- 
work for the routinization of the process of conflict. In industry, these 
include collective-bargaining bodies of many kinds as well as systems 

66 Changes Since Marx 

of conciliation, mediation, and arbitration. In politics, legislative 
bodies and courts of law serve similar functions/^ All these forms 
help to convert strikes and civil wars from an exclusive weapon of 
conflict to an ultima ratio of the contending parties. 

Such forms of conflict regulation proper have been supplemented, 
in most industrial societies, by changes in the structure of the industrial 
enterprise and of the state which aim at reducing the intensity of con- 
flict. The establishment of shop stewards and factory committees, 
and the participation of workers in industrial management, as well 
as certain rules and customs relating to the rights of opposition par- 
ties," are changes of this kind. There is an element of ideological mis- 
conception in many of these institutions. That is, though these new 
institutions may be designed to eliminate conflict, they may, indeed, 
actually increase the violence of conflict by redirecting it. In any case, 
these structural changes in those social organizations that generate 
conflict show the acceptance by society of conflicting interests, with 
which society has attempted to cope by institutionalizing them. 

The institutionalization of class conflict implies its continued exist- 
ence. But institutionalized class conflict is nevertheless far removed 
from the ruthless and absolute class struggle visualized by Marx. It 
is quite probable that most contemporary industrial societies have 
ceased to be capitalist societies. If this is true, it has happened not be- 
cause they were unable to cope with the contradictions and conflicts 
generated by the structure of capitalist society, but, more likely, pre- 
cisely because they were able to cope with their conflicts. Like many 
of his contemporaries — if with rather different evaluative accents^- 
Marx was so struck by the dynamics of early industrial conflict that he 
believed its satisfactory settlement utterly impossible except by a revo- 
lution. However, also like many of his contemporaries, Marx was 
mistaken. Some feared and others wanted the revolution, but both 
fear and hope were equally unfounded. Nobody can, of course, ever 

■^® For some discussion of the sociological aspects of these types of conflict regu- 
lation, and of the institutionalization of class conflict in general, see, apart from 
Geiger's worlc, the book by W. E. Moore (157) and the articles by Lockwood (86) 
and Kerr (84). A systematic, though more abstract, exposition of the regulation of 
social conflict will be given below, in Chapter VI. 

^^ By contrast to changes in the structure of the enterprise, those in political insti- 
tutions are very varied in different countries. They include, however, the participa- 
tion of opposition members in decision-making bodies, especially in matters of defense, 
the rights of higher civil servants to be retained even if they are not in the majority 
party, the custom to appoint a member of the opposition as speaker of parliament, and 
the like. 

Changes Since Marx 67 

be sure that a given pattern of conflict regulation will always prove 
successful. There are still strikes, and for all we know they will con- 
tinue to occur. But it has proved possible for industrial society to get 
along with the clashes of interest arising from its industrial and politi- 
cal structure — and it has proved possible for interest groups to get 
along with industrial society. Instead of a battlefield, the scene of 
group conflict has become a kind of market in which relatively autono- 
mous forces contend according to certain rules of the game, by virtue 
of which nobody is a permanent winner or loser. This course of devel- 
opment must naturally be bitter for the orthodox and the dogmatic, 
but theirs is the kind of bitterness which makes liberal minds rejoice. 


Not every industrial society is a capitalist society, and one of the 
differences between the time of Marx and ours might be described 
as the supersedence of capitalism. But while there are a number of 
significant differences between industrial societies of a hundred years 
ago and today, there are also similarities. Some of these are obvious. 
There are machines and factories, workers and entrepreneurs, wages 
and profits today as there were a hundred years ago. The historical 
traditions of the countries with which we are here concerned have 
been added to, but not changed. Many of the specific cultural and 
social features of capitalist societies in Europe and North America 
have survived the changes of the last century almost untouched. How- 
ever, we are not concerned here with a general and comprehensive 
account of industrial societies and their development. Within the 
more limited scope of the present study, the culture and history of 
societies is but a background before which specific structures and con- 
flicts unfold, and in turning to those elements of social structure that 
remained unchanged in the past century we shall again confine our 
discussion to aspects relevant to the problem of class. 

Every society has a structure of institutions, groups, and roles. 
But every society is also, as Durkheim put it, a "moral society," that 
is, it entails a set of norms and values which live both in the minds of 
its citizens and in the patterns of their social relations. One of the para- 
doxes of the social history of industrial societies is that the structure 
of their institutions has in the past century changed in many respects, 
while their values have merely advanced, but not changed. This is 
but a seeming contradiction, because one can show that most of the 
changes discussed in this chapter occurred within a framework which 

68 Changes Since Marx 

remained intact, and that the values of rationality, achievement, and 
equality belong to this framework rather than to its changing details. 

Thus, "economic rationalism," the value of purposeful economic 
activity oriented toward the maximization of gains, has never left 
industrial societies. As Max Weber has shown, this value preceded 
capitalism and, indeed, constitutes one of the factors which account 
for the emergence of capitalism. But its dynamics were not exhausted 
once capitalism was under way. Within capitalism, the impulse to 
organize economic activity more "rationally" remained a stubborn 
tendency. When, towards the end of the nineteenth century, many 
countries experienced what has sometimes been called a "second in- 
dustrial revolution," the value of rationality stood behind it: at about 
this time, the "extensive" increase of industrial production was re- 
placed by an "intensive" increase, i.e., by a "more rational" organiza- 
tion of existing resources. "Scientific management" and even "social 
engineering" were part of this trend. It was in the course of this de- 
velopment that ever larger concentrations of production enterprises 
proved imperative — enterprises so large that few individuals could 
provide the capital to establish them, so that some new form of owner- 
ship became desirable. At the same time, the usefulness of human 
skills in terms of the maximization of gains was rediscovered. It 
proved more "rational" to carry on production with experienced and 
well-trained workers than with mere "hands." Finally, rationality 
and bureaucracy are never very far apart; in one sense the rapidly 
growing demand for clerks and office employees, accountants and 
supervisors, statisticians and submanagers is but a concomitant of the 
"rational" organization of enterprises. Thus, the changes that led to 
the supersedence of capitalist society were the reflection of values on 
which this type of social and economic organization was based. Eco- 
nomic "rationalism" seems a value characteristic of all industrial so- 
cieties, not merely of their capitalist variant.^" 

Essentially the same holds for the value of achievement, i.e., the 
central place accorded to individual capacity, effort, and success in 
industrial societies. As with gains and profits, the content of success 
may change in the course of history, but achievement has remained, 

'° I do not think that the replacement of the "protestant ethic" by a "social ethic" 
(Whyte) contradicts this argument. If by "rationalism" we understand purposeful 
activity oriented to maximizing gains, it remains open whether these gains are entirely 
monetary in nature, or whether they include certain less tangible gains well compatible 
with a "social ethic." The opposite of "rationalism" is "traditionalism," and I see 
no evidence for a return of the latter. 

Changes Since Marx 69 

so far, the cardinal social virtue of the citizens of industrial societies. 
If anything, the recognition of achievement has progressed in the 
past century. Ever since the industrial revolution, ascriptive criteria 
of status have proved an obstacle to the systematic exploitation of re- 
sources, material and human. An industrial enterprise cannot afford 
to rely on the social origin of its members in the sphere either of man- 
agement or of labor j to carry on its tasks it needs above all capable, 
well-trained people. From the enterprise, this premium on achieve- 
ment has spread to the whole society. For this reason, industrial so- 
cieties need a minimum — if not a maximum — of social mobility; and 
for this reason also, the educational institutions have grown, in indus- 
trial societies, into the place of agents of role allocation. 

Like rationality and achievement, equality is also a value which 
characterizes all industrial societies, capitalist or otherwise. We have 
seen that legal equality is not merely compatible with, but conditional 
for the capitalist class system. Even Marx realized, despite his scorn 
for the "purely formal" ideology of equality, that "this one condition 
comprehendsa world history" (12, 1, p. 178). But we have also seen 
that "this one condition" has not remained confined to equality before 
the law, and its extension to other spheres was probably no accident. 
Inequalities of status, legal and social, obstruct the formation of ra- 
tional organizations in which the place of individuals is determined 
by achievement. Although the realization of equality remained piti- 
fully imperfect in capitalist societies, it was a part of the program of 
these as it is of their successors. 

There is only one point at which the changes in the structure of 
industrial societies since Marx might appear to have affected the 
values of these societies. It is often said that modern societies are 
always "open" (Weber) or "adaptive" (Mayo). And in seeking for 
an explanation of such terms one may encounter statements like: "We 
have in fact passed beyond that stage of human organization in which 
effective communication and collaboration were secured by established 
routines of relationship" ( 1 54, p. 1 2) ." But, then, this kind of state- 
ment is simply a misunderstanding of social reality. As I have pointed 
out before, post-capitalist societies do in fact display effective routines 
of human cooperation (e.g., the institutionalization of class conflict), 
but this is not a reason for assuming a change in values. The fact that 
the values of industrial societies were realized incompletely in their 

^^ This formulation, by the way, is a typical example of a statement which, in 
its logical status, is not sociological in that it does not permit of empirical test. This 
in itself would be sufficient reason to be skeptical about its validity. 

70 Changes Since Marx 

earliest, capitalist, forms is not surprising, nor does it affect the im- 
portance of these values for both capitalist and post-capitalist phases 
of development. We can conclude with some confidence that the social 
changes discussed in this chapter have all been compatible with, and 
to some extent consequent of, certain social values which advanced 
industrial societies share with their capitalist precursors.^^ 

Apart from these social values, certain other elements of the social 
structure of industrial societies have also survived the changes of the 
past century and are therefore characteristic of capitalist and industrial 
societies alike. These structural features are so formal and general 
in nature that they may well be elements not only of industrial but 
also of all human societies. Yet, while this general problem need not 
concern us here, they have to be mentioned, for they both have a spe- 
cial bearing on class structure and conflict. I mean, here, the existence 
of social stratification and of authority relations. 

Although, as the preceding discussion has incidentally brought 
out, patterns of social stratification have undergone considerable 
changes in the course of the last hundred years, these changes have 
not affected the existence of a hierarchical differentiation of status, 
and there is no indication that hierarchies of socioeconomic status will 
disappear in the foreseeable future. There still are rewards to be dis- 
tributed by society, these are still desired by individuals, and the dis- 
tribution of desired rewards is still unequal. I think one can go one 
step further and add that, by and large, the criterion by which rewards 
are unequally distributed is still that of occupation. Many contem- 
porary sociologists believe that we have left the age of work and pro- 
duction and entered an era of leisure and consumption. But to the 
present day the extent of a person's leisure, as well as the level of his 
consumption, is entirely determined by his occupation and the rewards 
associated with it. Whatever criterion of social stratification one pre- 
fers, prestige or income, spending habits or styles of life, education 
or independence, they all lead back to occupation. Social stratification 
based on occupation is as such neither the whole nor a part of class 
structure, but it does constitute an element of inequality which has 

^^ A second conclusion that can and must be drawn from this analysis of changing 
structures within unchanged values has to do with the usefulness of the value concept 
for the analysis of social change. Despite their boundaries, the changes discussed in 
this chapter are by no means unimportant; yet it would be hard if not impossible to 
retrace them on the level of values. Thus it would appear that only long-term changes 
(lii<e the emergence of industrial society) can be fruitfully analyzed in terms of values; 
for the other changes, it seems best to analyze them in terms of structural elements. 

Changes Since Marx Ji 

persisted and which continues to influence the issues and patterns of 
social conflict. 

This is even more clearly the case with respect to relations of 
domination and subordination as they exist in social organizations of 
both capitalist and post-capitalist industrial societies. The assertion 
that there still are authority relations in industry, in the state, and in 
many other forms of social organization does not go unchallenged in 
contemporary sociology. In surveying recent theories of conflict in 
the following chapter, we shall encounter a number of authors who 
believe that power is a thing of the past. Short of anticipating this 
discussion and giving a precise definition of what is meant by power 
and authority, I would here merely assert this: in post-capitalist as 
in capitalist industrial enterprises there are some whose task it is to 
control the actions of others and issue commands, and others who have 
to allow themselves to be controlled and who have to obey. Today 
as a hundred years ago there are governments, parliaments, and courts 
the members of which are entitled to make decisions that affect the 
lives of many citizens, and there are citizens who can protest and shift 
their vote but who have to abide by the law. Insofar as either of these 
relations can be described as one of authority, I would claim that rela- 
tions of domination and subordination have persisted throughout the 
changes of the past century. Again, I believe that we can go even 
further. The authority exercised in both capitalist and post-capitalist 
society is of the same type j it is, in Weber's terms, "rational authority" 
based "on the belief in the legality of institutionalized norms and the 
right of command on the part of those invested with authority by these 
norms" (33^^, p. 124). From this condition many others, including 
the necessity of bureaucratic administration, follow. But these are 
based above all on the fundamental social inequality of authority 
which may be mitigated by its "rational" character, but that neverthe- 
less pervades the structure of all industrial societies and provides both 
the determinant and the substance of most conflicts and clashes. 


Some Recent Theories of Class Conflict in 
Modem Societies 


Our account of changes in the structure of industrial societies since 
Marx has been largely descriptive, although I have tried to weave 
a number of threads of development into some main cords. Descrip- 
tion, however, is but the basis of scientific knowledge, not its proper 
substance. I am not sure whether Merton's statement that ". . . in 
building the mansion of sociology during the last decades, theorists 
and empiricists have learned to work together" (214, p. 102) was not 
a little premature when he made it in 1 948. Merton may well be right 
in his claim that empirical research can serve to inspire, transform, 
redirect, and clarify theories, apart from testing them. But it is still 
necessary to emphasize that fact-finding is not in itself explanation, 
and that therefore "empirical research" cannot replace "theoretical 

With respect to Marx's theories of class formation and class con- 
flict the overwhelming majority of social scientists have failed to pro- 
ceed beyond the level of description. There is a growing number of 
studies about occupational prestige and social mobility. A library of 
social psychological and sociological monographs has been written 
about the "new middle class." There are numerous investigations 
into trends of social stratification, egalitarian or otherwise. Research 
in industrial sociology often touches upon problems of industrial con- 
flict and the authority structure of the enterprise. But the authors of 
all these studies would do well to call them sociographic rather than 
sociological investigations. There is no lack of data, but cut bono} Of 
what use is the remark found in many a research report that its results 
rrfute one thesis or another thesis of Marx's? What is the use of re- 

■^ Of course, the reverse must also be emphasized: theoretical insight cannot re- 
place empirical research; but then, this misconception is very much rarer today than 
the one alluded to in the text. 

Recent Theories 73 

futing one or the other thesis, if the progress of science stops at this 
early point? It is certainly of some moment to refute theories and 
hypotheses derived from them. But it is more important to super- 
sede, on the basis of such refutations, the theories themselves, and to 
put better and more satisfactory ones in their place. From this point 
of view all the authors whose approaches we shall discuss in this chap- 
ter have advanced far beyond the so-called pure empiricists. 

Theory is "the net which we throw out in order to catch ^the 
world' — to rationalize, explain, and dominate it" (Popper, 218, 
p. 26). In this sense, a theory is "more" than a particular hypothesis, 
and a hypothesis is "more" than a concept or category. For Marx, 
"class" is a category. Its substance can be found by observation and 
described by the category. Thus the statement that "the proletariat 
is a class" is descriptive in that it subsumes an empirical datum under 
a concept. But the statement that the proletariat is involved in a con- 
flict with the bourgeoisie which can only be solved by a revolution is 
a hypothesis. It is derived from a theory, from which other hypoth- 
eses are also derived j we have called it Marx's theory of class conflict. 
According to this theory, every society generates in its structure con- 
flicting classes which develop in a certain way and the conflict of which 
eventually leads to structural upheavals. Thus a theory is a general 
"point of view" which structures an area of facts and transforms it 
into an ordered context. As scientists, it is not our task to "select only 
such facts as confirm the theory and, as it were, repeat it; the method 
of science is rather to look out for facts which refute the theory" (Pop- 
per, 219, II, p. 260). Facts refute theories by proving false one or 
several hypotheses necessarily following from them." In this sense, 
we can conclude that Marx's theory of class conflict — insofar as it is 
a scientific theory and not untestable speculation — has been refuted 
by the "facts" discussed in the preceding chapter. 

But the refutation of theories, although "the vehicle of scientific 
progress" (Popper), is not its substance. Refutation of old theories 
makes sense only if it becomes the point of departure for the develop- 
ment of new theories. This is what is meant here by the supersedence 
of a theory. Such supersedence may assume several different forms. 
In the case of Marx's theory of class it may mean that we restrict the 
validity of this theory to a single, limited type of social organiza- 
tion, such as capitalism. Alternatively, it may involve reformulating 

^ Properly speaking it is not "facts," but "statements of fact" that refute hypoth- 
eses. For a more detailed discussion of these methodological problems, see the quoted 
works of Popper. 

74 Recent Theories 

Marx's theory by modifying some of its elements. Finally, we may 
decide to dismiss this theory altogether and replace it by an entirely 
new theory of social conflict. However, one of these paths will have 
to be chosen if we want to advance our knowledge of social conflict 
and change. In the present chapter, we shall discuss some recent at- 
tempts to supersede Marx's theory in the light of historical facts and 
sociological insights. It will emerge from this discussion that none of 
these approaches has substituted a satisfactory new theory for that of 
Marx — that is, a theory which survives the test of observed data. It 
will then be our task to explore new paths and to hope that this ex- 
ploration will contribute to the development of a useful theory of 
social conflict. But before we embark on this long and at times some- 
what tedious journey, it is unfortunately necessary to free the central 
category of our analysis, the concept of class, of the jumble of mis- 
conceptions which obscures its meaning today, and to try and reinstate 
it in its proper place. 


The history of the concept of class in sociology is surely one of 
the most extreme illustrations of the inability of sociologists to achieve 
a minimum of consensus even in the modest business of terminological 
decisions.^ Only recently, Geiger, Lipset and Bendix, and some others 
have unraveled the conceptual confusion a little and begun to inaugu- 
rate a rather more rigid use of the category of class. It is not our 
intention here to discuss all the versions and perversions of the con- 
cept of class. That would be a task for the sociologist of knowledge j 
indeed, it has been done, with varying success and often not without 
additional confusion, since Mombert (23) and Sorokin (30), who 
counted thirty-two variations of the concept as early as thirty years 
ago, by Geiger (91, 46), Marshall (58), Cox (40), Pfautz (97), 
Lipset and Bendix i^SS)-) Croner (129), and others. For curiosity's 
sake, then, rather than in order to attempt a comprehensive survey, 
I shall quote here a small selection of definitions of the much abused 
concept of class: 

Class "is a force that unites into groups people who diff^er from one 
another, by overriding the diff^erences between them" (Marshall, 57, 
p. 114). 

"Class, as distinguished from stratum, can well be regarded as a psy- 

' I have discussed the conceptual problems raised here more extensively in an 
essay (41). 

Recent Theories 75 

chological phenomenon in the fullest sense of the term. That is, a man's 
class is a part of his ego, a feeling on his part of belongingness to some- 
thing j an identification with something larger than himself" (Centers, 
38, p. 27). 

"We shall then mean by a social class any portion of a community 
which is marked off from the rest, not by limitations arising out of lan- 
guage, locality, function, or specialization, but primarily by social status" 
(Maclver,211,p. 167). 

"According to the point of view here advanced, social classes . . . are 
social groups determined by three factors, namely, ( 1 ) similar social con- 
ditions, (2) similar social status, (3) similar social values" (Croner, 129, 
p. 185). 

"By class is meant two or more orders of people who are believed to be, 
and are accordingly ranked by the members of the community, in socially 
superior and inferior positions" (Warner and Lunt, 100, p. 82). 

Whoever reads these definitions may well be tempted to regard 
sociology as rather a frivolous discipline. Indeed, theories can neither 
be formulated nor refuted on the basis of these definitions, some of 
which are plainly bare of substance, others too profuse, and all far 
removed from the original purpose of the concept of class. Attempts 
have been made to classify the multitude of existing concepts of class 
itself. Usually, "objective" and "subjective" definitions of class are 
distinguished; Pfautz has added to this the distinction between "ex- 
ternal objective," and "internal objective" approaches (97). But 
such attempts achieve the opposite of their purpose, if this may be 
suspected in a clarification of the concept. We are indeed faced with 
an alternative: either we renounce the discredited term "class" alto- 
gether and endeavor to find a less ambiguous set of terms, or we reject 
radically all definitions which depart from the original, i.e., Marxian 
heuristic purpose, and return to this source. 

In a historical discipline it is always difficult to take a "protestant" 
view and ignore traditions. While the attempt to reinstate the concept 
of class in its original meaning is methodologically unobjectionable, 
we must also ask whether it is pragmatically feasible. There are legiti- 
mate doubts about the latter aspect, and to some extent I share these 
doubts. As a tentative and reversible decision, however, I propose 
to retain the term "class" and contrast it with others for which alterna- 
tive terms have already begun to become accepted. The criterion by 
which this distinction seems possible is the heuristic purpose of the 
categories in question. From this point of view, we shall have to reject 
all definitions of class as a category of social stratification whether 

76 Recent Theories 

they be "internal objective," "external objective," or "subjective." 
Wherever classes are defined by factors which permit the construction 
of a hierarchical continuum, they are wrongly defined 5 i.e., the term 
has been applied wrongly. Status, ranking by others, self -ranking, 
style of life, similar economic conditions, and income level are all 
factors which define social strata but not social classes. However one 
may interpret, extend, or improve Marx, classes in his sense are clearly 
not layers in a hierarchical system of strata differentiated by gradual 
distinctions. Rather, "the analysis of social class is concerned with an 
assessment of the chances that common economic conditions and com- 
mon experiences of a group will lead to organized action" (Lipset and 
Bendix, sS-> P- 2,48).* Class is always a category for purposes of the 
analysis of the dynamics of social conflict and its structural roots, and 
as such it has to be separated strictly from stratum as a category for 
purposes of describing hierarchical systems at a given point of time. 
This statement of the meaning of the concept of class (or, rather, 
of what it cannot properly mean) automatically excludes from the 
present investigation a large number of sociological works which 
overtly deal with "classes." For example, Warner's six-"class" theory 
(75, 100) — if it deserves the name of theory at all — constitutes 
neither a refutation nor a supersession of Marx's theory of class. It 
cannot be conceived as such, for, using the terms correctly, Warner 
should have referred to his groups of equal rank not as "upper- 
upper class y^ "lower-upper class j** etc., but as "upper-upper Stratum,* 
"lower-upper stratum,** etc. In the following discussion of recent 
theories, we shall therefore confine ourselves to three types of ap- 
proaches: those which are explicitly based on what, with a deliberate 
overstatement, I shall now call the correct concept of class 5 those 
whose authors assert that they are based on the correct concept of class ; 
and those which, although they are not based on this concept, are 
formulated in such a way as to suggest that they might genuinely 
supersede Marx's theory of class. Obviously, even those post-Marx- 
ian theories of class which conform with these standards are too numer- 
ous all to be included in a brief discussion. A comprehensive sum- 
mary and critique of these theories would require a lengthy book in 
itself. I have therefore selected, from a multitude of works, those 
which are comparatively recent and which are consequently not them- 

* At a later stage, I shall raise some objections to the exclusive reference to "eco- 
nomic conditions" in statements of this kind; they are nevertheless true to the heuristic 
purpose of the concept of class. 

Recent Theories 77 

selves accompanied by a host of Interpretations and critiques/ Fur- 
thermore, I have tried to choose examples which are characteristic of 
entire schools of thought. In this chapter, I have also left out purely- 
conceptual discussions, some of which will be mentioned later, and I 
have concentrated on analyses of modern societies in terms of class 
structure and class conflict. Finally, my selection has been guided by 
the fruitfulness of approaches for a new sociological theory of con- 
flict, the formulation of which constitutes the ultimate aim of the 
present study. 

(Nemchinov, Djilas) 

Evidently, those are faced by the most difiicult problems of analy- 
sis who, by a philosophical or political decision, are bound to main- 
tain Marx's theory in all essential features as an instrument of 
sociological explanation. One need hardly mention that this kind of 
decision makes fruitful scientific research and analysis impossible. 
Dogmatic insistence on scientific theories always mars the progress 
of knowledge. Quite possibly, this is one of the reasons why many 
of the analyses presented by Soviet scholars are so sterile and naive 
that in non-Communist countries they would hardly be acceptable 
for publication in scholarly journals or by serious publishers. Thus 
the brief study by Nemchinov (64), a member of the Soviet Academy 
of Sciences, is distinguished from the mass of these analyses not by its 
profoundness or thoroughness, but by its subject matter which Marx- 
ist social scientists'' have been extremely hesitant to deal with, namely 
"changes in the class structure of the population of the Soviet Union." 

To begin with, Nemchinov recalls the original definition of class 
and distinguishes it from the notion of social stratification, with which 
the analysis of class has in his opinion been confused in the West in 
an impermissible manner. "Thus, the objective criteria for the de- 
termination of social classes are the position of the society member in 
the occupation and the character of his income, determined by the pre- 

' This criterion excludes the whole Marxist literature. While a discussion of 
Marxist thinking on class from Kautsky to Lukacs and further would certainly promise 
many a useful result, it seemed to me an unnecessary ballast in a book that is primarily 
concerned not with Marx but with social conflict. 

® Perhaps it would be more precise to speak here of "Marxist-Leninist" or "ortho- 
dox Marxist" authors, for among Marxians, i.e., scholars who profess a scientific 
(which means last, not least, critical) relation to Marx, there have been many serious 
studies of class. See, e.g., the work of Djilas discussed below. 

78 Recent Theories 

vailing form of property and the type of productive relations in which 
individual members stand to other members under the particular 
system of social labor" (p. 14)/ There is but one interesting and 
potentially consequential point at which Nemchinov supplements 
Marx's theory. He separates, both conceptually and empirically, 
property and power, and he states that in bourgeois society "it is be- 
yond any doubt that to a greatest extent property relations provide 
an all-round control over the living conditions of the workers," 
whereas "common ownership of the means of production in the USSR 
makes it impossible to transform private income into a source of 
power" (p.. 1 3). This does not, of course, contradict the position of 
Marx, for whom effective private property, and not property in gen- 
eral, was the determinant of class structure, but it nevertheless opens 
up unexpected vistas. If property can, but must not, convey power, 
a curious alternative emerges of which Nemchinov has chosen the 
dogmatic, untenable side, whereas Djilas, as we shall presently see, 
chose to apply Marx's theory plausibly, although critically. Nem- 
chinov does not hesitate to state his "theoretical position" quite un- 
ambiguously. "The great historical changes of the twentieth century 
and those of the past centuries [ ! ] have confirmed the scientific ob- 
jectivity, truth and validity of the theory of social classes formulated 
by Marx and later developed and worked out by Lenin" (p. 3). It 
is nice to know that Lenin's "objectivity, truth and validity" was as- 
sured even before he was born. But if we ignore for the moment the 
fact that this statement is a confession rather than a proposition, we 
must ask : In what sense could it possibly be valid ? How does a theory 
based on such beliefs cope with the changes which industrial societies 
have undergone since Marx? 

Nemchinov takes a stand only with respect to two of the changes 
we discussed in the preceding chapter: to the problems of property 
and of the "middle class." Moreover as to the first of these, he has 
no more to offer than a few dogmatic assertions. "The material basis 
of bourgeois hegemony lies in the private ownership of the means 
of production. . . . The workers in a capitalist society do not pos- 
sess the means of production and are obliged to sell their labour to 
the employer, the owner of the means of production" (p. 12). This 
is but an inferior repetition of Marx and ignores completely the 

^ Nemchinov's essay, as well as the paper by his colleague P. N. Fedoseyev quoted 
below, was originally submitted to the Third World Congress of Sociology in English. 
All quotations are from these English texts. It is evident, however, that the translators 
at the disposal of the Soviet Academy of Sciences are rather inferior to those employed 
by the Soviet government and the Politbureau. 

Recent Theories 79 

phenomenon of the separation of ownership and control. However, 
Nemchinov's colleague Fedoseyev (45) settled this problem in ac- 
cordance with the party line, although I have to confess that I find 
his approach more amusing than impressive. Fedoseyev thinks that 
"it is not difficult to understand that the replacement of the indi- 
vidual capitalist owners by corporative monopolistic capital does not 
mean at all that the capitalist class disappears. The place of the 
individual owners of the capitalist enterprises is taken by the par- 
ticipants of the monopolistic unions, by the magnates of financial 
capital, millionaires and multimillionaires, who exploit the working 
people through the system of trusts, joint-stock companies, banks and 
state-capitalist enterprises" (45, p. 267). Had Fedoseyev read his 
Marx more thoroughly, he would have known that "the place of the 
individual owners" has been taken by at least two sets of people, of 
which one, the managers, can hardly be described as magnates and 
multimillionaires. But perhaps such plain facts are too insignificant 
for the imagination of Soviet thinkers. 

Not unexpectedly, Nemchinov and Fedoseyev are essentially 
agreed with respect to the "middle class." "The problem of the 
so-called middle classes plays a great role in the theory of social 
classes. Bourgeois sociologists assume that homogeneous societies can 
be achieved as a result of the growth and development of the middle 
classes" (64, p. 6). So far so good. But what is the "true" place of 
the middle class? "History has, however, shown," says Nemchinov 
(p. 6), "that this social group is economically unstable and subject 
to class disintegration." Fedoseyev goes even further. According to 
him, the middle strata "have declined not only in their specific share 
in production, but also in their numbers" (p. 268). Evidently, both 
authors refer here to the "old middle classes" of peasants, artisans, 
and shopkeepers. As to the "new middle class," Nemchinov confines 
himself to criticizing Clark's distinction of "primary," "secondary," 
and "tertiary" industries and asserting that the latter are "undoubt- 
edly the continuation of the process of production and must come 
under social production" (p. 8), and that therefore the old classes 
exist in these as in older forms of production.* 

There is, however, one aspect of Nemchinov's analysis which leads 

^ Nemchinov's disinclination to accept Clark's distinctions seems to me a charac- 
teristic illustration of the "sterility" and "naivete" of orthodox Marxist analyses: 
many concepts which are both useful and necessary for descriptive purposes are simply 
denied, as if such denials could conjure away the realities described by such concepts. 
This kind of deliberate primitiveness of the scientific instrumentarium makes com- 
munication difficult and differentiated insights virtually impossible. 

8o Recent Theories 

beyond such stereotyped repetitions of dogma. He deals at some 
length with the position of specialists, professional people, and higher 
technical and administrative employees. Nemchinov refers to this 
group as the "intelligentsia," or "intellectuals." "The intellectuals 
are an intermediate social group, the class character of which is de- 
termined by the prevailing method of social production. Under capi- 
talism, intellectuals stem mostly from the propertied classes and are 
closely connected with their own class. Among them there are also 
representatives of the workers' intelligentsia, closely connected with 
the peasantry and the working class. In socialist society intellectuals 
are mostly drafted from the workers and collective farmers and are 
closely connected with the working masses. Nevertheless the intelli- 
gentsia cannot be considered, under capitalism as well as under com- 
munism, as a special middle class. They are only an intermediate 
social group, existing along with the basic social classes" (pp. 9-10). 
Elsewhere, Nemchinov describes this group as an "intermediate social 

Here we find indeed something like a precarious theoretical at- 
tempt to explain a new phenomenon of social development with the 
categories of the old theory of class. However, this attempt is, in the 
first place, based on erroneous statements of fact. With respect to 
the Soviet Union, we still know comparatively little about the social 
origin of the "intelligentsia." But in "capitalist societies" of the 
present, this group is by no means solely, or even mainly, recruited 
from the "propertied classes." If we look at contemporary England 
(1949) as an example, we find that of all men in the status catego- 
ries I, 2, and 3 of D. V. Glass's study (107) — i.e., all higher salaried 
employees, and people in managerial and professional occupations — 
24.3 per cent (category i), 40.9 per cent (category 2), and 62.1 
per cent (category 3), respectively, were recruited from strata which 
Nemchinov would count among the "working class" (108, p. 183). 
Nemchinov has bracketed out the phenomenon of social mobility in 
advanced industrial societies, and his theory therefore explains at best 
one aspect of the social changes that have occurred since Marx. 

But does it really explain this aspect? It is Nemchinov's thesis 
that the "intelligentsia" is in principle a stratum which is neutral from 
the point of view of class, and which does not therefore constitute 
an autonomous force in class conflicts. Were it not for Nemchinov's 
definition of this group, this might sound rather like A. Weber's and 
Mannheim's theory of the "free-floating" or "socially unattached 
intelligentsia." But we must remember that Nemchinov includes in 

Recent Theories 8i 

the "intelligentsia" apart from "intellectuals" proper the "techni- 
cians employed in various branches of industry" as well as "admin- 
istrative workers," which latter groups presumably constitute the ma- 
jority of this strange stratum. It includes, indeed — as soon as Nem- 
chinov refers to Soviet society — government ministers, party func- 
tionaries, and the managers of industrial enterprises. In other words, 
Nemchinov claims that the incumbents of positions of authority in 
state and industry, and their bureaucracies, neither are a social class 
themselves nor provide a reason to modify Marx^s theory of class 
at any point. We shall see presently that the first of these claims is 
a matter of dogma rather than of Marxian analysis; for on this point 
Djilas's analysis seems perfectly convincing. The second of Nem- 
chinov^s claims, however, is manifestly false and leads its author into 
interesting contradictions. If it is true that the bureaucratic managers 
of state and industry do not form a class, and if it is further true that 
contemporary Western societies are still class societies in Marx's sense, 
then it follows that mere property (of shares, for example) without 
control can be the basis and determinant of a class. This consequence 
would undoubtedly modify Marx's theory, and it would reveal Nem- 
chinov as a representative of the "vulgar mind" which "commutes 
class differences into ^differences in the size of the purse' " (Marx, 5, 
p. 466). If, on the other hand, the incumbents of leading bureau- 
cratic positions do form a class in Western societies, it is hard to see 
why in the Soviet Union they should be a mere "stratum," an out- 
growth of the working class. Nemchinov's venture in social theory 
manages to repeat Marx's theory up to a point, but it falls victim also 
to the changed patterns of advanced industrial society. 

Where Soviet social science refers to the West, it is almost in- 
variably unconvincing, but also often colorful and amusing. "The 
abyss between labour and capital, between working people and ex- 
ploiters is not filled, but deepened, for the profits of the monopolists 
grow, whereas the share of the masses of the population in the national 
income steadily falls. The growth of technology leads to the inten- 
sification of labour and to greater exploitation of workers. This is 
the source of the aggravation of social contradictions and the basis 
of continuous class struggle in the capitalist countries" (Fedoseyev, 
45, p. 268). However, where Soviet social science refers to Com- 
munist countries, it is not only unconvincing but also remarkably 
barren and boring. Nemchinov reiterates in detail Stalin's theory of 
"nonantagonistic classes" — without of course mentioning, in 1956, its 
author. "The modern class structure of Soviet society has been re- 

82 Recent Theories 

fleeted in the Soviet Constitution of 1936. At present Soviet society 
consists of two basic classes, the working class and the collective 
farmers,^ as well as a social stratum, the intelligentsia" (p. 22). That 
the mere attempt to extend the concept of class in this fashion bears 
witness to the "inadequacy of the classical Marxist-Leninist concep- 
tion of class," Nemchinov could have learned from his less dogmatic 
Polish colleague Ossowski {^G^i^ p. 24). That it conceals rather than 
reveals and explains the realities of Soviet society seems evident, and 
has been convincingly demonstrated by Djilas's analysis of the "new 
class" (44). 

Djilas's work is of interest in our discussion, because its author 
applies a fairly strict Marxian approach to the analysis of Communist 
societies. Its subject matter is precisely that odd "intelligentsia" of 
managers, party secretaries, and bureaucrats which — if Nemchinov's 
study is at all symptomatic — appears to have puzzled Soviet scholars 
for some time. Only, Djilas cuts deeper than Nemchinov. He is not 
content with the problematic construction of an "intermediate social 
stratum" which belongs neither here nor there, but instead explicitly 
calls this "intelligentsia" a new ruling class with its own very special 
social characteristics. "It is the bureaucracy which formally uses, ad- 
ministers, and controls both nationalized and socialized property as 
well as the entire life of society. The role of the bureaucracy in so- 
ciety, i.e., monopolistic administration and control of national income 
and national goods, consigns it to a special privileged position. . . . 
Ownership is nothing other than the right of profit and control. If 
one defines class benefits by this right, the Communist states have 
seen, in the final analysis, the origin of a new form of ownership or 
of a new ruling and exploiting class" (44, p. '}^'^^. 

The main peculiarity of the bureaucratic ruling class of Commu- 
nist societies is, according to Djilas, that it has not grown spontane- 
ously like other classes in history, but has been the deliberate creation 
of a party elite. It follows from this fact that the rule of the new 
class is more brutal and all-embracing than that of any other class 
in history. But otherwise this "new class of owners and exploiters" 
(p. 54) resembles earlier classes in many ways. Its domination is 
based on ownership, for, as Djilas argues at great length, "collective 
ownership" is but a facade behind which "this new class, the bureauc- 
racy, or more accurately the political bureaucracy" (p. 38), exercises 

^ See 64, p. 15. "The workers and collective farmers are friendly classes in the 
USSR, as their material interests are not opposed. . . ." 

Recent Theories 83 

its control. Coupled with its "monopolistic ownership" the new class 
has "totalitarian authority," that is, its power extends over all spheres 
of life. Like earlier classes, it tries to legitimize its rule by an elabo- 
rate ideology. Finally, Djilas hopes and predicts that, like other 
classes in history, this new class will also be overthrown and displaced 
in a revolutionary action of the oppressed. 

Djilas naturally does not include the contemporary Western scene 
in his analysis, so that we cannot infer from his work how he would 
try to modify Marx's theory to fit the Western case. But it is an 
important fact to remember that his analysis shows that a fairly strict 
Marxian analysis might still be applicable to Communist countries. 
This conclusion may mean either of two things. It is conceivable that 
a Marxian theory of class conflict applies fully only to countries 
undergoing the process of industrialization. But it is also possible 
that we can infer from Djilas's analysis that some kind of Marxian 
theory is of more general usefulness and applicability. I have delib- 
erately referred to Djilas's approach as a fairly strict Marxian analy- 
sis, and as so^ne kind of Marxian theory. There is, in this approach, 
a slight shift of emphasis at a crucial juncture of Marx's theory which 
may prove essential to the problem of the applicability of this theory. 
This shift of emphasis concerns the relation between ownership, or 
property, and power. 

I think it can be shown that Djilas is not unambiguous when he 
speaks of the determinants of the new class. On the one hand, he 
says, "As in other owning classes, the proof that it is a special class 
lies in its ownership and its special relations to other classes" (p. 44), 
or, "The specific characteristic of this new class is its collective own- 
ership" (p. 54). On the other hand, one can find a statement like, 
"Today power is both the means and the goal of Communists, in 
order that they may maintain their privileges and ownership. But 
since these are special forms of power and ownership, it is only through 
power itself that ownership can be exercised. Power is an end in itself 
and the essence of contemporary Communism" (p. 169). There is 
at least a hint here that it is ultimately not the ownership of the means 
of production that determines a class, but that this very ownership is 
only a special case of a more general social force, power. While Marx, 
as we saw, subordinates relations of authority to those of property, 
Djilas seems inclined to subordinate ownership to power. The prob- 
lem of the relationship of these two factors will recur time and again 
in the considerations of this study j and it wilj, I hope, become in- 
creasingly clear that in explaining social conflicts and changes, there 

84 Recent Theories 

Is greater promise in an approach that follows Djilas than follows 



It was of course not merely by Nemchinov, and not only in Com- 
munist countries, that the attempt has been made to save Marx's 
theory of class as a principle of explanation for advanced industrial 
societies. Two lines of argument are characteristic of such attempts 
at rescue. On the one hand, certain new facts of social development, 
such as the separation of ownership and control or social mobility, are 
either denied or explained away. This was done, for example, by 
Kuczynski, whose attempts to "prove" the "pauperization" of the 
proletariat in "capitalist" societies despite a continued rise in real 
wages (152) has been exposed by Geiger in all its untenability and 
even ridiculousness (46, pp. 60 ff.). Obviously, a dogmatic approach 
of this kind does not advance our knowledge of society. On the other 
hand, seemingly unimportant elements of Marx's theory are modi- 
fied by such "revisionists" without their realizing how the original 
theory is being transformed in their hands. Thus, Renner introduced 
in his early attempt at a systematic formulation of Marx's theories 
(26) an interesting additional distinction. He argues that recent social 
development has complicated class structure, especially the bourgeoi- 
sie. "It puts alongside the capitalist who owns and functions the other 
one who owns but does not function. . . . What is more: It also 
produces the non-capitalist who exercises capitalist functions, who 
therefore does not own but functions as a capitalist" (26, p. 375). 
For us this is by now a familiar line of argument. But if one regards 
in this way the merely controlling manager as a "capitalist," i.e., as 
the founder of a class, then class conflict is separated from its Marx- 
ian root, private property. The assertion loses its meaning that a 
["socialist"] society based on communal property is a classless society. 
Thus, an argument of this kind cannot save Marx's theory either. 
Saving Marx's theories is of course in any case a highly doubtful 
undertaking, in which a social science should take good care not to 
become involved. But in his early writings Renner has neither wanted 
nor clearly seen the necessity of superseding Marx's theory of prop- 
erty and social class^°j it was Schumpeter who first carried through 

^° By contrast to the later Renner, whose work will be discussed below. The 
attempt to supersede Marx on the basis of his own principles and yet retain the political 
impetus of "scientific socialism" expresses itself even more clearly in the work of Ren- 
ner than in that of Bernstein. 

Recent Theories 85 

the conscious separation of Marx's theory of the development of 
property from his theory of class conflict. 

Schumpeter poses Marx's question of the "economic law of 
development of capitalist society" anew. Can capitalism survive? 
And, can socialism work? Both "capitalism" and "socialism" are for 
Schumpeter categories that describe economic systems — in particular, 
property relations. In this sense, Schumpeter agrees with Marx's 
conclusion that the economic order of capitalism is bound to break 
down and give rise to a new economic constitution which is socialist 
in that it is based on communal or, as Schumpeter prefers to say, state 
ownership. This is a "necessary" process — not, however, in a Marx- 
Hegelian sense of immanent laws of historical development, but as 
a scientific "statement about the tendencies present in an observable 
pattern," which does not "tell us what will happen to the pattern but 
only what would happen if they continued to act as they have been 
acting in the time interval covered by our observation and if no other 
factors intruded" (73, p. 61). 

If, up to this point, his analysis is quite compatible with that of 
Marx, Schumpeter now goes his own way. Socialism is, as he says, 
"culturally indeterminate J . . . in fact, . . . a society may be fully 
and truly socialist and yet be led by an absolute ruler or be organ- 
ized in the most democratic of all possible waysj it may be aristocratic 
or proletarian j it may be a theocracy and hierarchic or atheist or in- 
different as to religion j it may be much more strictly disciplined than 
men are in a modern army or completely lacking in discipline" (73, 
pp. 170-71). It may be — Schumpeter does not say so at this point, 
but it corresponds to his conception — a class society or not. Although 
this, too, has to be inferred from Schumpeter's text, it would seem 
that for him socialism betrays many traits of a class society 5 for the 
work of workers "would remain substantially what it is" (p. 203) j 
"there will be plenty to fight about in socialist society" (p. 213), for 
which "the problem of bureaucratic management" is not the least 
cause (pp. 205 ff.). 

For Schumpeter, this kind of analysis becomes possible by a com- 
plete and deliberate departure from Marx's theory of class. "Marx 
wished to define capitalism by the same trait that also defines his class 
division. A little reflection will convince the reader that this is not 
a necessary or natural thing to do. In fact it was a bold stroke of 
analytic strategy which linked the fate of the class phenomenon with 
the fate of capitalism in such a way that socialism, which in reality 
has nothing to do with the presence or absence of social classes, be- 

86 Recent Theories 

came, by definition, the only possible kind of classless society, except- 
ing primitive groups. This ingenious tautology could not equally 
well have been secured by any definitions of classes and of capitalism 
other than those chosen by Marx — the definition by private owner- 
ship of means of production" (p. 19). 

It is evident from this passage that although Schumpeter would 
like to retain the concept of class as an instrument of analysis, he 
considers it a useful instrument only if it is freed of its Marxian con- 
nection with private property. Thus we must ask: What kind of a 
theory of class promises, in the opinion of Schumpeter, a more satis- 
factory analysis of the processes of development of industrial so- 
cieties? The answer is, Schumpeter presents neither a theory nor an 
analysis of class of his own. One might almost be tempted to believe 
that his (potentially correct) thesis of the "cultural indeterminate- 
ness" of economic systems has led him to the (false) conclusion that 
cultural and social processes cannot be explained systematically and 
therefore do not permit scientific theories. However, this assumption 
is contradicted by the fact that in an earlier work (27) Schumpeter 
has in fact dealt extensively with the phenomenon of class and de- 
veloped the rudiments of an original theory. Unfortunately — and 
I suspect that this is the ultimate root of the uselessness of Schum- 
peter's analysis for the explanation of class conflict and its develop- 
ment — this theory was based on a concept of class which we have 
explicitly dismissed above, namely, that "class structure is an arrange- 
ment of individuals according to their social rank as it varies by groups, 
ultimately according to difFerences of ability j even more, it is based 
on the fact of the institutionalization of such rank once achieved" 
(27, p. 205). Schumpeter himself defines the area of problems to 
which his "classes" belong when elsewhere he refers to "social strata 
or groups" and adds, "As is known, we call them social classes" (99, 
p. 216). Schumpeter's classes are indeed not classes but strata. 

At one point of his analysis Schumpeter has begun to approach 
a genuine supersedence of Marx's theory of class. He has shown that 
the necessary connection between the existence of classes and the 
existence of private property asserted by Marx is empirically untena- 
ble. Thus he has opened up the possibility — later seized by Djilas — 
of analyzing also, with the instrument of class theory, societies in 
which private property no longer functions or exists. But Schumpeter 
has, by diluting the concept of class into that of stratum, made it 
impossible for himself to make use of this critical decision. Sering, 
who follows Schumpeter in dissociating the concepts of class and pri- 

Recent Theories 87 

vate property, goes further here, if he states with respect to the 
transition, from capitalist to modern industrial societies that "the new 
social order does not emerge without classes" (74, p. 205). However, 
Sering's analysis (which is not, on the whole, very original) pre- 
supposes another theory which under the name of the "managerial 
revolution" has become a widely used slogan. 

(Burnham, Croner) 

Although James Burnham refers with suspicious frequency to 
science in general and the scientific nature of his analyses in particu- 
lar, his theories are not very clear and his analyses not very acute/^ 
He wants to establish the fact of a social revolution, and yet he says 
(140, p. 7), "I am going to assume further . . . that the present is 
in fact a period of social revolution." At one point, the managers, 
dramatis fersonae of his revolution, are described rather specifically 
as "the operating executives, production managers, plant superintend- 
ents, and their associates" (p. 82) j but later this concept suddenly 
includes not merely "production executives" and "directing engi- 
neers," but also quite generally "administrators," "propaganda spe- 
cialists," and "technocrats" (p. 281). He proposes to demonstrate 
that the managers are everywhere about to assume power, but many 
of his formulations lend support to the suspicion that he simply calls 
managers those who are powerful. It would not be difficult to dis- 
miss Burnham's theory almost oifhand if all we wanted to prove was 
that it is formulated in a most unsatisfactory manner and therefore 
not, as such, a useful instrument of sociological analysis. However, 
I propose to avoid this simple path and try to explicate in the work 
of Burnham the elements of a theory that might supersede Marx's. 

Burnham, too, starts off by discussing the tendency to supersede 
capitalism which is inherent in modern economic and social develop- 
ment. The new fact which he is trying to explain is the phenomenon 
described above as the separation of ownership and control. Burn- 
ham. does not like this expression. "The truth is that, whatever its 
legal merits, the concept of 'the separation of ownership and controP 
has no sociological or historical meaning. Ownership means control \ 
if there is no control, then there is no ownership. ... If ownership 
and control are in reality separated, then ownership has changed hands 

^^ He is also not the first author to present a theory of managerial rule, as Schel- 
sky has pointed out with reference to Veblen, Rathenau, and others (cf.' 164). 

88 Recent Theories 

to the Control,' and the separated ownership is a meaningless fiction" 
(p. 92). Terminological (and confused) as this argument may 
appear, it is nevertheless basic for Burnham's theory. Because he 
defines property ownership^^ by control, he can describe the transi- 
tion of control over industrial means of production from owner- 
managers to mere managers as a process which leads to the super- 
sedence of capitalism without affecting class conflict in its Marxian 
definition as relating to property ownership. 

"In all complex societies so far, there is a particular, and rela- 
tively small, group of men that controls the chief instruments of 
production" (p. 59). In capitalist society this was the capitalist class. 
But in the society toward which we tend today "the managers will, 
in fact, have achieved social dominance, will be the ruling class in 
society" (p. 72). In this society, legal property lies with the state, 
but factual control over the means of production is exercised by the 
managers who, in this sense, are their owners. And "the instruments 
of production are the seat of social domination j who controls them, 
in fact not in name, controls society, for they are the means whereby 
society lives" (p. 125). Like their capitalist predecessors, the new 
"owners" of the means of production, the managers, are confronted 
with the propertyless class of workers. "The managerial economy is 
in actuality the basis for a new kind of exploiting, class society" (p. 
122). Indeed, exploitation is worse here than in capitalist society. 
On the one hand, the working class becomes increasingly homogene- 
ous, for "in comparison with the organization of industry in the period 
prior to modern mass production, the individual tasks . . . require 
relatively less skill and training on the part of the individual worker" 
(p. 77). There is a wide gap between largely unskilled workers and 
highly specialized managers. On the other hand, in bureaucratized 
managerial societies, even the dubious liberty to sell his labor is taken 
from the worker and he is forced to work (cf. p. 130). We are still 
in a "transition period," but before long all industrial societies will 
be managerial class societies. 

Clearly, Burnham's theory, even if it proves useful, can at best 
only partially supersede Marx's theory of class. It is based on one 

^^ An accident of language has probably contributed to no small extent to make 
Burnham's theory possible. The German word Eigentum comprises both the passive 
"property" and the active "ownership," and whereas a separation of property and con- 
trol is a plausible occurrence, one of ownership and control seems indeed somewhat 
doubtful, i.e., the element of control is indeed somewhere implied in the term owner- 
ship, though not in property. 

Recent Theories 89 

new fact, the separation of ownership and control. We can therefore 
trace it back to Renner's somewhat liberal interpretation of Marx, 
or even to Marx's own analysis of the phenomenon of joint-stock 
companies. (Incidentally, Burnham's identification of the managerial 
rule with state property in the means of production indicates an im- 
portant limiting case, but it is really quite unnecessary j the separa- 
tion of legal and effective ownership would be fully sufficient for his 
theory.) Other tendencies of development of industrial societies, 
such as the growth of social mobility, do not contradict Burnham's 
theory but remain unexplained by it. The theory of managers merely 
modifies one aspect of Marx's approach. 

There are, moreover, a number of tendencies which directly con- 
tradict Burnham's theory. Burnham follows Marx in asserting the 
increasing homogeneity of the working class, brought about by virtue 
of the leveling of their skills and qualifications. We have seen that 
this thesis is false. The working class has not in fact become more 
uniform with respect either to its skill structure or to differentiations 
of income and prestige j there are, rather, tendencies to the opposite. 
Burnham further refers to an increase in the intensity of class con- 
flict in managerial society. This assertion, too, is hardly compatible 
with what we have described as the institutionalization of class con- 
flict, nor, for that matter, does it agree with the fact that the legiti- 
macy of managerial authority rests to a considerable extent on the 
consensus of those subjected to this authority. Also, as we have seen, 
there are many tendencies toward decreasing the intensity of class 
conflict and toward enlarging the sphere of institutionally regulated 
clashes. Here, too, Burnham is in error. 

However, the crucial and most problematic aspect of Burnham's 
theory is that he blindly follows Marx in identifying economic with 
political power and domination. "The sources of wealth and power 
are the basic instruments of production, these are to be directed by 
the managers J and the managers are, then, to be the ruling class" 
(pp. 158 f.). Not surprisingly, Burnham's critics pounced on this 
observation. "The real question is," states Bendix (126, pp. 119 f.), 
"whether (i) men who control the policies of industries, govern- 
ment, labor unions, farm groups, etc., constitute a cohesive group 
owing to this common characteristic; or (2) the ideas and policies of 
the so-called managerial group differ in any respect from those of 
the older type of entrepreneur." These are both empirical questions 
which cannot be answered by dogmas or definitions. But Burnham 
neither proved nor could he prove that in real societies the property- 

90 Recent Theories 

less managers of industry are in any sense identical with the ruling 
groups of the state. Instead it appears that "either by violating the 
principle of identity or by taking the term 'manager' as an emblematic 
slogan to mean those in power, Burnham exploits the facts concern- 
ing the growth of bureaucratic structures for his own thesis" (Gerth 
and Mills, 148, p. 173). 

Burnham tried to abridge the pedestrian process of scientific dis- 
covery, and for this he had to pay. The facts which he wants to con- 
nect systematically in his theory originate in the sphere of industrial 
production. But the theory which he presents claims to be a valid 
explanation of the development of the entire society. Between its 
claim and its basis there gapes an abyss of implicit assumptions and 
dogmatic assertions. The value of Burnham's theory lies in its con- 
sequence that the class structure of the industrial enterprise is based 
on control and not on legal ownership of the means of production. ^^ 
With this hypothesis, the separation of ownership and control becomes 
a phenomenon which is fundamentally irrelevant for class conflict. 
But the problem of political conflicts in advanced industrial societies is 
logically and empirically independent of that of the class structure 
of the enterprise and must therefore be investigated separately. The 
simple assertion that the means of production are the "seat of social 
domination" explains little, even if it is repeated a hundred times. 

In fact, many sociologists have tried for more than thirty years 
to find an answer to the second, more comprehensive question of the 
political structure of advanced industrial societies and its class basis. 
Usually, approaches to this problem are focused on the phenomenon 
of bureaucracy, i.e., on those parts of the "new middle class" which 
are more or less intimately connected with public administration. 
Burnham tries in vain to ridicule what he calls the "theory of the 
bureaucratic revolution," and to dismiss it with a few superficial argu- 
ments (pp. 278 ff.) . No great effort is needed to realize that the analy- 
ses of Max Weber and Michels, Bendix and Merton, and many 
others" have contributed more to our knowledge of the structure of 
contemporary societies than Burnham with his theory of the mana- 
gerial revolution. Yet neither Weber nor most of the subsequent 

^^ Although I would doubt whether it is sensible to express this approach with 
Burnham by an extension of the definition of property ownership and to hide thereby 
a substantial difTerence behind the verbal agreement with Marx. 

■^* For a good collection of the most important texts of the authors mentioned 
and of others, see the reader edited by Merton and others (136). 

Recent Theories 91 

analysts of bureaucracy were explicitly concerned with developing a 
theory which might lead to the supersedure of Marx's theory of class 
in the light of new facts. Rudiments of such a theory can first be 
found in the work of Geiger to which we shall turn presently, and, 
if on a much less ambitious level, in that of Fritz Croner ( 129). 

Like Burnham, Croner believes he is able to perceive in the devel- 
opment of industrial societies in the twentieth century a "social revo- 
lution." "Everywhere the social revolution with which we are dealing 
here has changed the face of society fundamentally. ... Its product 
and its bearer is a new social class: white collar" (p. 9). "Capitalist 
order created in its beginning the social space for a new class, the exist- 
ence of which rests on individual economic effort. The age of rational- 
ization has also created the social space for a new class: the class of 
salaried employees. . . . This process is the real cause why society 
today appears in an entirely different light than society fifty years ago: 
this social process is the real substance of the ^social revolution' of our 
time" (p. 246). If Croner speaks of a "real cause" and of the "bear- 
ers" of a social revolution, he evidently aims at an explanation of the 
changed structures of advanced industrial societies. However, he does 
not regard himself as an opponent or superseder of Marx. For him, 
"there is no systematic theory of class in the work of Marx" (p. 169), 
because Croner does not see, and much less realizes, the heuristic pur- 
pose of the analysis of social conflict or the analytical departure from 
antagonisms based on the distribution of property or power. Thus, 
he says "class" but he means "stratum" when he refers to "similar 
economic conditions," "similar social status," and "similar social val- 
ues" as determinants (p. 185). On the strength of this fact, we should 
have to exclude Croner's contribution from this survey as being out- 
side its scope. Yet Croner has presented — perhaps unwillingly — 
with his "theory of delegation" and his account of the functions of 
white collar, an approach which might be understood as a contribution 
to our discussion and which I shall regard as such. 

The "new" fact with which Croner is concerned is the spectacular 
growth of the "new middle class" since 1890. Croner explains this 
fact as a subdivision of entrepreneurial functions in industry and of 
leading positions in the state, which had become necessary in the 
course of rationalization. "The explanation for the special social 
position of salaried employees can be found in the fact that their work 
tasks used to be entrepreneurial tasks" (p. 36). "What we . . . 
have said about the division of entrepreneurial 'power' and the re- 

92 Recent Theories 

suiting emergence of certain 'services' in the economy, can be applied 
also to the emergence of the civil 'service.' Here of course it is not 
the entrepreneur, but the highest chief of state, e.g., the king, who 
delegates certain tasks which he has so far carried out himself to men 
of his confidence who then 'represent' the king in their fields" (p. 37) . 
Croner does not mention the separation of ownership and control, 
but his "theory of delegation" presupposes this process to a certain 
extent. His approach also includes the explanation of one aspect of 
social mobility,^^ if he says: "The salaried employees are a class with 
considerable variations of income, influence and prestige. . . . When 
the salaried employee enters his first job, he may have a host of super- 
ordinates. ... At the end of his career he may have hundreds and 
thousands of subordinates. . . . But this whole career occurs within 
one and the same class" (p. 195). 

One could derive from these statements a theory which would 
argue approximately as follows. By the separation of ownership and 
control and the rationalization of industrial and political administra- 
tion, a new class of bureaucrats has emerged, the functions of which 
are subdivided authority functions. In this sense, the new class is a 
ruling class, and even the only ruling class. It is mobile; there is 
within it a continuous upward movement; and the class is by no means 
uniform as a social stratum. But by virtue of its share in (delegated) 
authority it is in conflict with all other groups in society. 

To repeat this, Croner himself does not formulate a theory of 
this kind. The concept of ruling class plays no part in his investiga- 
tion. Yet Croner's analysis contains certain elements of such a theory 
— elements which can of course also be found in the works of Burn- 
ham and of Weber and Michels, and which deserve our attention. The 
historical trends which we have emphasized may suggest certain re- 
strictions of this theory (such as of the notions of salaried employee 
or bureaucracy), but they do not refute it. However, it is evident that 
in speaking of bureaucracy as a ruling class one implies a concept of 
class which has little to do with that of Marx. Whether some exten- 
sion or modification of managerial and bureaucratic theories might be 
considered a useful explanation of post-capitalist societies and their 
conflicts, we cannot decide before we have settled the problem of 
method, substance, and limits of class analysis in sociological research. 

^^ Namely, intragcneration mobilit}'. The formulation quoted here might, how- 
ever, also be thought to account for the multitude of social origins of members of a 
class in a society in which intergeneration mobility is institutionalized. 

Recent Theories 93 


Both Burnham's and Croner's works suffer from the peculiar in- 
consequence of books that were quite obviously written to be best- 
sellers. In profoundness of thought and significance of analysis they 
compare badly with two late, posthumously published essays of the 
Austrian Karl Renner (71), whose sociological work — if it is known 
at all — is often underestimated. Renner refers to himself as a "Marx- 
ist." But this epithet means for him that although he wants to retain 
the "Marxian method" he is quite prepared to apply this "method"^^ 
in a critical and unprejudiced fashion to the new realities of post- 
capitalist industrial society. For it is for Renner "crystal-clear that 
the factual substratum, the social substructure, has changed com- 
pletely in the last hundred years" (p. 214). In fact, Renner begins 
his analysis of class structure in the essays under discussion, contrary 
to his program, with a critique of Marx's "method," of the theory 
of class itself. He adds to the statement that for Marx private prop- 
erty was the basis of class formation. "But obviously there have been 
domination and exploitation of other kinds in history, and in my opin- 
ion the Marxian school, although it has not overlooked and has in 
fact occasionally analyzed these, has failed to investigate system- 
atically and balance with each other all historical and possible relations 
of authority" (p. 89). This sentence implies more than its cautious 
formulation betrays. It indicates that Renner — much like Djilas, 
Schumpeter, Burnham, and others — is trying to free the concept of 
class from its definitional tie with private property and to apply it 
more generally to all relations of domination and subjection. Ren- 
ner's further considerations confirm this conclusion. His catalogue 
of possible forms of domination — "stratocracy," "capitalism," "the- 
ocracy," "graphocracy," "bureaucracy," etc. — is at the same time sup- 
posed to be a catalogue of possible forms of class conflict. 

By this modification of Marx's approach, Renner can, following 
Burnham, analyze the separation of ownership and control without 
assuming that it involves the elimination of class conflict. "Besides 
the capitalist who has lost his function, there stands . . . the func- 
tionary who has lost his capital : a social character mask of great future 

^® Renner makes nowhere quite clear what he means by this "method." But his 
use of the notions of a "factual substratum" and a "normative superstructure" occa- 
sionally suggests that he subscribes to a considerably modified "historical materialism" 
which for him probably represents the "Marxian method." 

94 Recent Theories 

significance" (p. 182). Contrary to Burnham, however, Renner re- 
gards the managers as "a stratum which, at least for the time being, 
is politically anonymous," and for which "a general solidarity of 
interest and ideology ... is no longer given" (p. 215). This con- 
clusion is made possible for Renner because he regards neither the 
"managerial revolution" nor the "white-collar revolution" as an iso- 
lated phenomenon which in itself characterizes advanced industrial 
society j these two social changes he consistently connects. For Renner 
the emergence of a managerial stratum is merely part of a general 
development in the course of which a new class emerges — the "service 

Before Croner, Renner developed a kind of "theory of delega- 
tion." In post-capitalist society, "the functions of capitalists appear 
subdivided in a steadily growing number of salaried employees of the 
very highest and of high and of lower rank. . . . These new aids are 
neither capitalists nor workers, they are not owners of capital, they 
do not create value by their work, but they do control values created 
by others" (p. 119). Renner calls this stratum the "service class." 
It has fashioned itself on the model of public civil service and has 
been transformed from a caste into a class. Although it participates 
in authority, it does not exercise absolute authority but is subject to 
the norms and laws of society. 

There is a further point at which Renner goes beyond Burnham 
and Croner. He does not confine himself to the analysis of changes of 
the ruling class j he also deals with changes in the position of the work- 
ing class. By its political and industrial achievements, the working 
class has today become a constituent element of society. "Today, the 
working class is no longer that incoherent sum of helpless individuals 
who are exposed to the storms of economic crises and the arbitrary 
rule of rulers as the desert sand is exposed to the elements. It is no 
longer the proletariat of 1 848, but a powerful, confident, well-organ- 
ized member of society. It is this member position which gives it 
power, and often more power and security than the possession of pri- 
vate wealth" (p. 211). By virtue of security thus guaranteed, the 
wage system based on effort and output has been replaced by a "liveli- 
hood system" based on needs. "Thus the working class and the service 
class have moved closer together" (p. 123). Capital is "controlled 
by a service class which obviously amalgamates more and more with 
the working class" (p. 226). 

For Renner, industrial society after the disappearance of the "full 
capitalist" is characterized by two classes: the service class and the 

Recent Theories 95 

working class. There still are, of course, the "finance capitalists," 
shareholders, and bankers, but Renner prophesies their impending 
end. Thus the basic question of class analysis has to be asked anew 
(p. 102), "Who or what has the real power today?" The "what" in 
the formulation of this question already betrays Renner's unortho- 
dox answer: working class and service class gradually amalgamate. 
Neither is really a ruling class. The ruling force is, rather, a "some- 
thing," something "objective" which Renner expresses by the notions 
of "norms" or "law." "As in a democratic state the citizen does not 
obey the person of the monarch but complies with the laws, as he does 
not serve the official but the office as the bearer of lawful mandates, 
so every employee now works according to the plan of the enterprise" 
(p. 102). The "general will of society" has replaced the "rule of a 
minority." "This general will defines the aim for society and thereby 
for the economy, and all functionaries pass over from the service of a 
master to the service of the whole. This general will is the law. . . . 
The exclusive rule of the law makes all forms of political illiberty 
impossible. The law creates the general order. Adaptation and sub- 
ordination by law do not create a state of illiberty." Nor does the 
execution of legal norms create a ruling class, for it is "a matter of 
institutions which are organized exclusively under economic and tech- 
nical aspects." "Economic democracy supplements political democ- 
racy (p. 227). 

This is, of course, Rousseau versus Marx — a contest that produces 
some fascinating results. But fascinating as they are, these results 
are not, perhaps, altogether convincing. Renner does not always make 
it quite clear whether his analysis refers to the present or to tendencies 
of the future. Yet he is evidently not concerned with long-term pre- 
dictions, much less with dreams of a classless society ä la Marx. From 
the separation of ownership and control, the extension of citizenship 
rights and equality, the institutionalization of class conflict, and the 
emergence of the "new middle class," Renner infers the formation 
of two large classes. But these are "nonantagonistic classes" or, rather, 
they blend into each other and tend to amalgamate. Renner's theory 
of class structure has many traits which appear well compatible with 
the changed realities of post-capitalist society and which it will there- 
fore be useful to bear in mind. But this cannot be said of the final step 
and climax of Renner's analysis, the postulate of a class society with- 
out class conflict. 

That classes can exist without entering into violent relations of 
conflict is a thesis not permitted by Marx's theory of class 5 neverthe- 

96 Recent Theories 

less, it might well be correct. It is at least conceivable that it corre- 
sponds to observable conditions. There is no reason to contest Renner's 
analysis merely because he asserts a class society without class struggle. 
What we have to reject, however, is the idea of a class society without 
class conflict^ because this notion is incompatible with the heuristic 
purpose of concept and theory of class. In arriving at his conclusion, 
Renner makes two significant mistakes. He confuses, first, the aspects 
of integration and conflict, and, second, the levels of institutions and 
norms. Instead of applying both to all societies, he looks at capitalist 
society in terms of conflict and institutions and at post-capitalist society 
in terms of integration and norms. Throughout his work, Renner 
recognizes a dualism of "factual substratum" and "normative super- 
structure" and considers both in any given society. Then, as he comes 
to deal with the new society of the present, he suddenly forgets this 
dualism and concentrates on the "normative superstructure" alone. 
"Who or what has the real power today?" he asks. He answers the 
"what" and neglects the "who." In fact, Renner should have asked 
in the beginning: "Who Ä«<i what have the real power today?" For 
it is always its institutional structures and its values that characterize 
a society — just as every society may and must be regarded from the 
point of view of its unity and coherence as well as from that of its 
inherent contradictions and conflicts. By forgetting the dichotomies 
of his earlier approach, Renner has been misled into an almost Stalin- 
ist notion of "nonantagonistic classes" in a happily integrated society. 
Yet his is an interesting mistake which will occupy us a great deal 
more in the present study. 

There is an empirical point also which must be held against Ren- 
ner's strange conclusion. Every society has its norms and laws, includ- 
ing what might be called ruling norms and laws. Norm is a very gen- 
eral category. Its relevance for class conflicts begins only when we 
ask which particular norms and laws are prevalent in a given society 
and which groups or aggregates of people either tend to enjoy privi- 
leges or suffer deprivations by virtue of the prevalent norms. The 
statement that the rule of the capitalists has been replaced by that of 
"the law" or "the general will" is remarkably meaningless, unless 
Renner wants to imply that in post-capitalist society power and its 
unequal distribution no longer exist. If, however, this is what he in- 
tended to say, there is little evidence to support his guess. There are 
many useful elements in Renner's analysis, including the assumption 
of the emergence of a service class vis-a-vis the working class, but his 
conclusions are disappointing in the light of general theoretical con- 
siderations as well as empirical evidence. 

Recent Theories 97 


All authors mentioned so far insist on the possibility of describing 
post-capitalist society, like its precursor, as a class society in an unmodi- 
fied or restricted Marxian sense. This premise is abandoned by Gei- 
ger in his study of the Class Society in the Melting-Pot (46). For 
Geiger, the old class society stands on the threshold of a new type of 
order, the structure of which can no longer be adequately compre- 
hended in the notion of class conflict. Significantly, the phenomenon 
of the "new middle class" — the "experts" and "bureaucrats" — is 
one of the maj or facts leading Geiger to this conclusion. Conceptually, 
Geiger prepares his conclusion by distinguishing "social stratum" as 
a general category from "class" as a "special case of social stratum," 
the case, namely, of a grouping "determined by the relations of pro- 
duction" (p. 35). This distinction is not purely terminological, and 
we shall have to question itj but this conceptual discussion will be 
postponed for the time being in favor of empirical critique. In any 
case, it is easy to renounce conceptual polemics against Geiger, since 
he himself has repeatedly stated his position with great clarity (20, 
46, 91), and has based his considerations on a fairly precise and in- 
sightful account of Marx's concept and theory of class. 

Geiger's "penetrating essay" (Marshall, 95, p. 13) is, despite 
its essayistic features and its restriction in evidence and argument, in 
its plan and execution perhaps the most ambitious attempt so far 
undertaken by a sociologist to master the changed reality of advanced 
industrial society in terms of a theory that is critically inspired by 
Marx. Geiger begins his study with a fairly elaborate critical account 
of Marx's concept of class and his "doctrine of the class society." He 
strongly emphasizes the "dynamic" purpose of class analysis and 
rejects objections against the model of two "dominant" classes result- 
ing from this purpose as "unfounded." There are, for him, well- 
founded objections against the application of the Marxian class model 
to post-capitalist society: the pauperization of the working class has 
not come true j the working class has neither become more uniform nor 
grown proportionately J the "old middle class" is still in evidence 
and has not become "proletarianized"; the capitalists are dying outj 
class consciousness has not increased. Before this background Geiger 
develops his central thesis. "The Marxist model of the industrial 
class society was presumably not inappropriate for the period of high 
capitalism" (p. 156). But even before this class structure "was able 
to penetrate the whole society, many other structural trends broke 

98 Recent Theories 

into the picture, deflected the stratification of 'capital and labor' and 
obscured it" (p. 157). Today, we are living in a process of transition 
from class "stratification" to a new type of "stratification." "The 
transition of society from one type of stratification to another one 
means . . . that hitherto subordinate lines of stratification become 
dominant, and hitherto dominant ones fade into the background" (p. 
153 f.). In this sense, the class conflict based on relations of produc- 
tion has given way in advanced industrial societies to certain "new 

Evidently, the proof of Geiger's theories lies in what he calls 
"new lines" and, furthermore, in the method that leads him to dis- 
cover these trends. Geiger proceeds cautiously and empirically. "The 
class society of Marxist coloring is obviously in retreat. Nobody can 
tell as yet with assurance which direction development will take. But 
we can today point out a number of competing trends and, with due 
caution, venture some suggestions as to the weight and force with 
which these will contribute to the future formation of society" (p. 
158). Geiger discusses in some detail five such tendencies. If we 
neglect the conflict between town and country (which presumably 
figures so largely in Geiger's book because it was first published in and 
for Denmark), four remain: the tendency of independent political 
action on the part of the "old" and "new" middle strata, that of the 
increasing importance of the consumer status, that of a conflict between 
all participants of production and the "mere consumers," and that of 
what Geiger calls the "rule of the experts." Geiger does not deal 
with the phenomenon of social mobility. But otherwise he explains 
the majority of conflicts arising out of new developments so convinc- 
ingly that one is inclined to forgive him his unjust and superfluous 
conclusion, "Thus the doctrine of Karl Marx is nothing but the 
anti-ideology corresponding to the liberal social reality of his time" 
(p. 228). 

Geiger's account of the four significant "new lines" of develop- 
ment is not without interest for our discussion. ( i ) As to the position 
of the middle strata in advanced industrial societies he notes a strange 
paradox. The "old middle stratum" of independent artisans and 
shopkeepers was engaged simultaneously in an economic struggle with 
large-scale capitalism and in an ideological struggle with the prole- 
tariat. In this two-sided conflict it is today supported by the "new 
middle stratum." Both resent the notion of class structure itself; 
they fight it in the name of an estate ideology which, in Germany, led 
them to support national socialism. "A class denies indignantly that 

Recent Theories 99 

it is a class, and it carries on a bitter class struggle against reality and 
idea of the class struggle" (p. i68). (2) The redistribution and 
equalization of income has cut through the class fronts. Besides the 
social status based on a man's "position as consumer," "the Marxian 
relation of production fades" (p. 175). "Most workers have become 
petty bourgeois in their purchasing power and spending habits" (p. 
176). (3) Thanks to the "institutionalization of class conflict," the 
social partners of industry have come closer to one another. Both 
derive a profit from every increase in production j in that sense one 
can "speak of a proximity of interest among capital and labor within 
urban industrial society. . . . The victims are those strata of society 
which one could describe as mere consumers, i.e., those who have no 
immediate share in the production and sale of material goods. . . . 
Poverty increases with the distance of an income-earner from com- 
modity production" (p. 194). (4) Which is the new "ruling stra- 
tum"? According to Geiger, Burnham has to some extent correctly 
diagnosed a tendency of development. But his assertion of the politi- 
cal rule of the economic managers is nonsensical j rather, "political 
bureaucratism absorbs the economy itself" (p. 217). "Thus the rule 
of the managers of private economy has not come about. In a cen- 
trally planned economic and social order power belongs to the eco- 
nomic officials, and if the name bureaucracy sounds too forbidding, 
one may replace it by 'rule of the experts' " (p. 220). 

The empirical references of Geiger's analyses are as plausible as 
most of his arguments. But the question remains to be asked. What 
is the theoretical position that has enabled Geiger to supersede Marx? 
In what sense do Geiger's "new lines" provide an analysis of advanced 
industrial societies that is equivalent to Marx's analysis of the capital- 
ist class society? Once again, the answer is disappointing. Geiger 
has not superseded Marx. His theoretical position is restricted to what 
is really just the formal statement that some societies display a kind 
of "stratification" different from that based on relations of produc- 
tion.^^ His analysis can be condensed to the single statement that the 
transition from early to advanced industrial societies puts hitherto 
subordinate lines of "stratification" in the place of class conflict. 

^^ Although Geiger always speaks of "stratification," he is in fact dealing with 
social conflicts. Yet he cannot be spared the charge of having obscured the essentially 
sound distinction between "stratum" and "class" by not distinguishing sufficientlv 
stricdy the difTerent "heuristic purposes" (this, too, one of Geiger's concepts!) asso- 
ciated with "stratum" and "class" and by falsely subordinating that of the concept of 
class to that of the concept of stratum. 

100 Recent Theories 

Popper has described scientific theories by the metaphor of the 
"searchlight." "What the searchlight makes visible will depend upon 
its position, upon our way of directing it, and upon its intensity, color, 
etc. J although it will, of course, also depend very largely upon the 
things illuminated by it" (219, II, p. 260). But always the search- 
light illuminates only a sector of reality, i.e., every theory is, as such, 
selective. It guides analysis to facts which are relevant for a particu- 
lar context and excludes others. This is precisely what Geiger's 
"theory" does not achieve, and could not achieve, because it is not 
really a theory. For this reason, Geiger's "new lines" consist in fact 
of a number of barely connected descriptions concerned partly with 
elements of social conflict (mere consumers against participants of 
industry), partly with changing relations of domination (rule of ex- 
perts), partly simply with general social trends (consumer status). 
There is no point at which these lines converge j they have not 
emerged in the selective beam of the searchlight of a theory but have 
instead been chosen arbitrarily from the infinite number of actual lines 
and trends of development. If they are nevertheless not without 
significance, this is a tribute to the good sense of a sociologist of Gei- 
ger's stature. Of the two aims of his investigation Geiger has achieved 
only one. He has shown that the model of a class society based on the 
relations of production is no longer applicable to the analysis of post- 
capitalist society. But he has not been able to show what must be sub- 
stituted for this model in order to render the new society accessible 
to sociological analysis. 

(Marshall, Schelsky) 

By several authors, one of Geiger's "new lines" has recently been 
made the point of departure of the attempt to penetrate the changed 
structure of post-capitalist society by means of sociological analysis: 
the increasing equalization of the social status of incumbents of differ- 
ent social positions, and in particular of the old classes. T. H. Mar- 
shall, for whom status emphasizes, in analogy to the legal meaning 
of the term, "the fact that expectations (of a normative kind) exist 
in the relevant social groups" (94, p. 13), has investigated this tend- 
ency at the example of the extension of the citizenship status. Accord- 
ing to his theory, the main features of which we have summarized 
above, the basic rights common to all citizens have been extended, 
during the last two centuries, to ever new spheres of social life. At 
first, legal status lost its differentiating force j later, political status 

Recent Theories lOi 

followed suit J and recent social development is characterized by a 
tendency to equalize the social position of all citizens with respect to 
the rights and privileges associated with this status. There are many 
indications of this trend, including equality of educational and occupa- 
tional opportunity, and the generalized right to a minimum income, 
to sickness benefits, and to old-age pensions, etc. "The basic human 
equality of membership . . . has been enriched with new substance 
and invested with a formidable array of rights. ... It has been 
clearly identified with the status of citizenship" (57, p. 9). Marshall 
does not confine himself, however, to an account of processes of social 
history, but goes on to pose the sociological problem, "Is it still true 
that basic equality, v/hen enriched in substance and embodied in the 
formal rights of citizenship, is consistent with the inequalities of social 
class?" (57, p. 9). His attempt to answer this question makes Mar- 
shall's analysis an essential contribution to our problem. 

To all appearances, equality rights associated with citizenship and 
class antagonisms are incompatible quantities. "Citizenship is a status 
bestowed on those who are full members of a community. All who 
possess the status are equal with respect to the rights and duties with 
which the status is endowed. . . . Social class, on the other hand, is 
a system of inequality. ... It is therefore reasonable to expect that 
the impact of citizenship on social class should take the form of a con- 
flict between opposing principles" (57, pp. 28 f.). In capitalist class 
society, however, this was the case only if one conceives of classes as 
groups endowed with different rights and privileges that are com- 
pletely closed off from each other. For classes in the Marxian sense, 
the principle of equal civil rights was, "on the contrary, necessary to 
the maintenance of that particular form of inequality" (p. 33) : legal 
equality was the basis of the (labor) contract, on which at the time 
inequalities of class were founded. Even the extension of citizenship 
rights to the political sphere did not seriously affect inequalities of 
class. It is only when the principle of universal basic rights is extended 
to the social sphere that the existence of classes becomes problematical. 
Of course, there still are inequalities. In a way, these are even caused 
by equal rights: "Through education in its relations with occupational 
structure, citizenship operates as an instrument of social stratification" 
(p. 67). But in a society in which these differences are not only re- 
duced to a minimum but also stripped of their ascriptive character, 
they cease to be productive of conflicts. "It may be that the inequali- 
ties permitted, and even molded, by citizenship do not any longer 
constitute class distinctions in the sense in which that term is used of 
past societies" (p. 75). In so far as class distinctions survive — and 

102 Recent Theories 

Marshall seems to believe that there is a limit beyond which inequali- 
ties cannot be abolished — they become "socially acceptable," lose their 
"economic function," and thereby their force as determinants of 
classes (see pp. 77 ff.). 

In a later essay (95), Marshall has supplemented this thesis by 
the assertion that classes may not have disappeared altogether, but 
have changed their character: they are no longer homogeneous inter- 
est groups whose unity is based on a common position in production 
and molded by common behavior patterns and ideologies, but "asso- 
ciations" formed temporarily in the occupational sphere "for the pur- 
suit of . . . common interests whenever they arise and with such 
degree of combination of groups as they demand" (p. 12). Such as- 
sociations are not as permanent and comprehensive as classes. "This 
might be described as a weakening of class in the Marxian sense, on 
the grounds that the operative interest groups are no longer deter- 
mined by the social relationship within the system of production, that 
is, primarily by property" (p. 13). 

Marshall refers at this point to Geiger's study. But by contrast 
to those of Geiger his theses do contain elements of a theory that might 
be thought to supersede that of Marx. Geiger has merely demon- 
strated that Marx's theory of class conflict is de facto insufficient for 
the analysis of post-capitalist societies, because it fails to account for 
certain new lines of social structure. Marshall suggests that the ex- 
tension of citizenship rights has penetrated the social sphere, and that 
these rights gradually lead to the elimination of all relevant inequali- 
ties, so that only such "class distinctions may survive . . . which have 
no appropriate economic function" (57, p. 77). If these suggestions 
are correct (which is open to dispute), then Marshall has shown that 
there can be no class conflicts in the Marxian sense in advanced indus- 
trial societies, because the structure of these societies is subject to a 
different law of development. If Marshall's theses are valid, the 
theory of an increasing institutionalization of an equal basic status of 
all citizens represents a genuine supersedure of Marx's theory. It 
explains both why capitalist society was a class society and why there 
are no classes in post-capitalist society. Before we examine the theo- 
retical and empirical basis of this theory, I shall supplement its formu- 
lation by an independently developed variant which places the empha- 
sis on slightly different evidence: Helmut Schelsky's thesis of the 
leveling of class society. 

So far, Schelsky's approach must be inferred from a number of 
scattered papers (cf. 72, 98, 164). In one of these, he explicitly 

Recent Theories 103 

poses what he calls (72, p. 62) the "naive, but nevertheless frequently 
heard question": "Have we still got a class society today?" "Class 
theory," the Marxian version of which Schelsky correctly describes 
(p. 61)) as concerned with the "decisive law of development and the 
dominant structure of society" was "a legitimate explanation of social 
reality. But is it still that?" To begin with, Schelsky replies rather 
cautiously, "Sociology will not be able to answer this general question 
with a plain Yes or No" (p. 62). In so far as "something which ex- 
isted once never disappears entirely in history, the question whether 
today there are still class structures has of course to be answered in 
the affirmative" (p. 63). But the answer is different if we ask whether 
class relations are still the dominant force in the structure and devel- 
opment of advanced industrial societies. Is, in other words, this so- 
ciety still characterized by one conflict of interests towering above all 
others and creating an "abyss between the classes"? "As a social scien- 
tist, one would have to answer this question today clearly in the nega- 
tive: in this sense we are at present no longer a class society. Why? 
Since the time which Marx had in mind, several social processes have 
occurred which have leveled-in and mitigated that abyss of class ten- 
sion, and at the same time new social structures and patterns have 
emerged which, more than the remnants of class conflict, must be 
regarded as the dominant structures of contemporary society that 
determine its course" (p. 64). In support of his thesis, Schelsky re- 
fers to the consequences of three parallel processes: (i) social mo- 
bility, (2) the leveling of styles of life, and (3) the inadequacy of 
surviving "ideologies" for the explanation of contemporary society. 
"Comprehensive and structurally far-reaching processes of up- 
ward and downward social mobility" have, according to Schelsky, 
"diminished class conflicts and leveled society into a very wide, com- 
paratively uniform social stratum" ( 72, p. 64). By these processes of 
mobility Schelsky seems to mean above all collective processes, such 
as the "collective rise of industrial labor," the "rise of . . . salaried 
employees into the new middle stratum," and the declassement par- 
ticularly of the "strata of the former bourgeoisie by property and edu- 
cation." Elsewhere, however, Schelsky also refers to the increase in 
the "mobility of single families," i.e., social mobility in the proper 
sense of the sociological concept, "which has broken up collective ties 
and solidarities of stratification to a large extent and has created an 
egoism of small groups and families as one of the essential social 
forces of our social constitution," so that "this highly mobile society 
cannot provide a lasting social status consciousness" ( 122, p. 5). 

104 Recent Theories 

The growth of collective and individual mobility is accompanied 
by a process of "leveling" social differences in the realm of income 
and prestige and above all in patterns of behavior and styles of life. 
This latter process, although emphasized also by Marshall, is central 
for Schelsky's conception. One could call this new and uniform "pat- 
tern of life . . . *petty bourgeois' or ^middle class,' if it were not for 
the fact that these terms lead to too many misunderstandings on ac- 
count of their class character" (72, p. 6$)- As a less ambiguous term 
Schelsky proposes that of the "leyeled middle class society." "This 
comparative leveling of formerly class- or stratum-determined pat- 
terns of behavior with respect to family life, occupational and educa- 
tional aspirations of children, the functions of living, consumption 
and support — indeed, of the cultural, political, and economic patterns 
of reaction in general — is perhaps at present the most dominating 
process in the dynamics of our modern society" (72, p. 6$). "It ap- 
pears that in place of class status the consumer position is becoming 
the central determinant of all patterns of behavior ... so that the 
negative process of leveling the class society would have to be de- 
scribed positively as the emergence of the highly industrialized lei- 
sure and consumer society" (pp. (>S f-)- 

At first sight, the trend of social mobility — which, after all, pre- 
supposes differences of position — and the leveling of styles of life 
appear to contradict each other. Indeed, "the paradox of our social 
change consists in the fact that the urge to rise on the social ladder 
has become universal at a time at which this ladder' has been com- 
pletely abolished or at least severely shortened. By reaching an extent 
that converts it into the fluid structure of society itself, social mobility 
contradicts the motives of individuals which are effective in it" (p. 
71). For this reason a sort of "false self-consciousness" emerges or, 
as Schelsky calls it, a "constitutional irreality of social self-conscious- 
ness" (98, p. 4) : because other explanations are lacking, men experi- 
ence society in categories which no longer hold. The need for security 
effects the "retention of antiquated notions of social status inherited 
from class or estate society" ( 72, p. 71). 

Schelsky does not assert that in advanced industrial societies there 
are no tensions or conflicts j indeed, he strongly objects to what he calls 
a "Utopian 'bourgeois' idea or hope of social harmony" (p. 66) . How- 
ever, the tensions peculiar to the new society are no longer class con- 
flicts. They are, rather, typically tensions "between the abstract orders 
of society and little primary groups" (231, p. 350), conflicts "of the 
individual or the immediate 'We' with the anonymous system of every 

Recent Theories 105 

kind of bureaucracy, on which one depends and by which one feels 
controlled, and even exploited" (72, p. 67). Elsewhere Schelsky 
claims in almost literal agreement with Marshall that "the conflict 
of organized representations of interests" has "replaced the conflict 
of the large, diffuse blocs of classes" (p. 68). 

This summary gives a rough outline of a kind of social theory the 
influence of which seems to grow steadily today in sociology as well as 
in public opinion. Marshall has formulated it in terms of an analysis 
of English social development, Schelsky with reference to (West) 
German society. We shall presently turn to American studies which 
resemble those of Marshall and Schelsky in many points. Thus it 
seems permissible to infer that a conception claims validity for all 
advanced industrial societies when its main arguments are as follows: 
( I ) the dynamics of post-capitalist society can no longer reasonably 
be described by a Marxian notion of class conflict j (2) the equalization 
of basic rights, conditions of life, and patterns of behavior has re- 
moved the basis of class differentiation j (3) moreover, an extraordi- 
nary intensification of social mobility makes the formation of lasting 
solidary groups impossible j (4) the (Marxian) theory of class has 
therefore lost its value as an instrument of explanation and must be 
replaced by a fundamentally different theory of social tensions on the 
basis of equal positions and situations. It will now be our task to con- 
sider whether these theses, and especially the second and the third of 
them, stand up to a critical examination in the light of empirical evi- 

It can indeed be shown, as has been indicated in the previous chap- 
ter, that a certain equalization of aspects of the situation of various 
social groups has taken place in the course of social development dur- 
ing the past century. Marshall and Schelsky plausibly emphasize two 
aspects of this development: the extension of equal basic rights and 
the leveling of differences of social status. On the other hand, there 
can be no doubt that this egalitarian trend has by no means advanced 
to the point of equality. It is certainly true that equal opportunity of 
education and upward mobility is a characteristic value orientation of 
industrial societies. But has this principle been realized yet? As D. 
V. Glass summarizes an empirical investigation, "Though between 
the two World Wars there was an unprecedented expansion in the op- 
portunities for secondary and university education, the discrepancies 
in educational opportunity as between individuals of different social 
origins had by no means been eliminated" (107, p. 16). 

The same holds for Germany, the United States, and other socie- 

io6 Recent Theories 

ties on a similar level of development. Even today legal and political 
citizenship rights are often restricted by social deprivations and disad- 
vantages. The right to proceed against others is of little moment for 
him who has not got the money to loose his case or even engage a 
lawyer. Differences not only of income and prestige, but also of the 
"consumer position" are even more apparent. The "comforts of civili- 
zation" are still distributed quite unequally, at least in the European 
countries to which Marshall's and Schelsky's analyses refer. Al- 
though in 1955 nine out of ten German households possessed an 
electric iron, only one out of ten had a washing machine or a refrig- 
erator, and only two out of ten a motor vehicle (229, pp. 27 f.). Al- 
though the cinema is open to everybody in his leisure time, there are 
considerable differences between those who make use of this possibility 
quite regularly (Germany, 1955: 46 per cent), rarely (30 per cent), 
or never (24 per cent). Although almost everybody listens to the 
radio (Germany, 1955: 92 per cent), not everybody listens to the 
same programs (cf . 229, pp. 59, 62, 67, etc.) . Of course, neither Mar- 
shall nor Schelsky claimed that "equality" or "leveling" are as yet 
completely realized; both refer to "tendencies" of social develop- 
ment; but, trivial as it may sound, it is necessary to emphasize time 
and again that so far the realization of these tendencies is extremely 
incomplete, and that the empirical validity of any theory based on 
these tendencies alone is therefore severely restricted. We have raised 
the question before, whether there might not be structural limits be- 
yond which the leveling of status symbols, rights, and situations can- 
not advance. It may well be that Marshall and Schelsky have been 
too fascinated with the leveling of traditional statuses and status sym- 
bols to realize that new, more subtle criteria are already taking their 
place. ^* 

However, this kind of objection in terms of empirical evidence 
cannot as such be regarded as a refutation of Marshall's and Schelsky's 
theories. There is no point in disputing the existence of leveling ele- 
ments of social status in modern societies. The really important ques- 
tions which we have to ask are of a different order: which spheres 
have been affected by this tendency? Are there any areas of social 
structure in which a tendency of this kind is not discernible? If we 
examine the theories of Marshall and Schelsky from the point of view 
of these questions, a peculiar fact becomes apparent. Schelsky's claim 

^® I am thinking here of makes of cars, kinds of leisure time activity, patterns of 
participation in culture, etc., in the sphere of statuses and their symbols, and of merely 
customary privileges of attendance and belongingness in the sphere of rights. 

Recent Theories 107 

of a tendency of leveling statuses, styles of life, and patterns of be- 
havior indicates above all a factual assimilation of status differences 
and their symbols. Marshall's notion of equal citizenship rights em- 
phasizes beyond that a shrinking of the sphere of possible social dif- 
ferences and, thereby, a lessening of possible sources of conflict. Both 
tendencies have certain obvious consequences for class conflict, in par- 
ticular for the substance of the opposing interests. But both tenden- 
cies leave entirely untouched one problem which is of crucial impor- 
tance for Marx's as for any other theory of class, namely, the problem 
of power or authority and its social distribution. 

Earlier we found that with respect to the social distribution of 
positions invested with authority a leveling tendency is hard to imag- 
ine and has certainly not occurred in modern societies. Probably, dif- 
ferences of legitimate authority are a basic fact of social structure j in 
any case, they are a fact systematically overlooked by Marshall and 
Schelsky. Their theories do not contain an answer to the problem of 
the position of the manager or bureaucrat, and, what is more, this 
problem is not even raised. The theory of leveling does not supersede 
Marx's theory of class but redirects analysis to different problems and 
different aspects of social structure on the silent or explicit assumption 
that in advanced industrial societies these aspects are in some sense 
more important and more dominating. Why they should be more im- 
portant, and above all why the distribution of authority should no 
longer be important — these are questions to which Marshall and 
Schelsky give no answer. For them, turning away from the problem 
of power is a thematic decision and not a result of analysis. The theory 
of leveling may not be wrong; but it is ultimately simply irrelevant 
for the problem of class. It deals with a different subject, that of 
social stratification, and it becomes objectionable, therefore, if the 
misleading attempt is made to pretend that it supersedes Marx or 
represents a comprehensive account of the dynamics of post-capitalist 

This objection of ours requires one slight modification. As we 
have seen, Marshall and Schelsky do not deny the existence of con- 
flicts and tensions in post-capitalist society. Schelsky in particular 

^^ Marshall and Schelsky can be accused of illegitimately applying a theory of 
social stratification to class analysis only in so far as both imfUcitly pass bv the proper 
area of class analysis. But this charge applies in full to some authors who have recently 
tried to prove that contemporary Western society is classless, whereas Eastern societv 
is a quasi-capitalist class society, and who have developed this argument by using two 
entirely different conceptual schemes for these two types of social order. 

io8 Recent Theories 

speaks of the "bureaucracy, on which one depends and by which one 
feels controlled, and even exploited." Elsewhere he refers to the 
"rule of the managers" (164) and even abandons the term "leveled 
middle-class society" in favor of "industrial-bureaucratic society" 
( 1 63, pp. 275 ff.) . But these conflicts and tensions have no systematic 
place in his theory. If one excludes the problem of power and au- 
thority from social analysis, one abandons the possibility to trace social 
conflicts back to structural conditions. They become essentially random 
phenomena, carried on by unpredictable chance groupings and related 
to fundamentally uncertain issues. For this is precisely what Marx's 
theory had achieved: to demonstrate the structural determinateness 
of social conflicts. An approach that throws overboard the intention 
and formal achievement of a theory with its particular, if empirically 
refuted, formulation does not supersede this theory but withdraws 
a sphere of knowledge from science in order to hand it over to the 
randomness of arbitrary opinion. 

Thus the theory under discussion is by no means what it purports 
to be, a supersedure of Marx's theory of class. Yet there is one ele- 
ment of it, the relevance of which for the problem of class we have 
to examine briefly. Marshall and particularly Schelsky place great 
emphasis on the phenomenon of social mobility in the sense of individ- 
ual movements up and down the status scale within and between gen- 
erations. Schelsky asserts that mobility has become "the fluid structure 
of society itself" and "has broken up collective ties and solidarities 
of stratification to a large extent." This assertion implies the hypoth- 
esis that classes lose their raison d^etre in a society in which belonging- 
ness to all nonprimary groupings assumes a merely temporary charac- 
ter. We shall have to return to this complex problem repeatedly. But 
the hypothesis of the impossibility of class formation in highly mobile 
societies contain an error which must be exposed at this point. 

In analyzing social structures of entire societies or individual insti- 
tutions, associations, and groups, a clear distinction is necessary be- 
tween social positions or roles and their agglomerations on the one 
hand, and the personnel of these units, the incumbents of such posi- 
tions, on the other hand. Social mobility constitutes in the first place 
a type of recruitment of the personnel of given positions. Social 
classes, however, are phenomena which at least potentially exist inde- 
pendent of the mode of recruitment and rate of fluctuation of their 
members. An industrial enterprise does not cease to exist if the annual 
rate of turnover of its workers amounts to lOO per cent or even 200 

Recent Theories 109 

per cent and more. In this sense, the degree of social mobility is as 
such irrelevant to the problem of the existence of classes. "Every 
class," Schumpeter remarks with a plausible metaphor, "resembles 
for the duration of its collective life ... a hotel or a bus which is 
always occupied, but always by different people" (27, p. 171). It is 
therefore false to assume that social classes and social mobility are 
as such incompatible. This is not to say, of course, that the increasing 
institutionalization of upward and downward mobility does not re- 
quire certain modifications of the theory of class. 


(Drucker, Mayo) 

It might appear strange if we conclude this survey of some recent 
theories of class conflict with a discussion of the work of two sociolo- 
gists in the analyses of whom the concept of class plays no part at all.^° 
Yet there is a certain logic in the sequence of theories discussed in this 
chapter. We began with a conception which at least pretends to be 
based on a strictly Marxian model. The theories of Djilas, Schum- 
peter, Burnham, Croner, and Renner involve increasingly consequen- 
tial modifications of Marx's approach. From these modifications it is 
only a step to Geiger's thesis that class structures have lost their 
dominant character, and further to Marshall's and Schelsky's as- 
sumption of the leveling of inequalities in post-capitalist society. The 
analyses of Drucker and Mayo — which, more than the others, may 
be regarded as arbitrarily selected examples for a widespread con- 
ception — perfect this line of analytical development by v/orking with 
entirely different categories. The problem of class conflict was at 
J east touched upon by all other authors mentioned in this survey j 
for Drucker and Mayo it no longer seems to exist. We shall have 
to find out what Drucker and Mayo have to offer instead of the 
Marxian theory of class, in so far as they deal with this aspect of 
social structure at all. 

The books by Drucker (144) and Mayo (154) with which we 
are here concerned have many features in common. Both authors 
believe — like Burnham and Croner — that modern social develop- 
ment involved a "revolution." "The world revolution of our time 

^° Both authors occasionally use the word "class" in the sense of "stratum." Even 
tJiis is rare, however, and furthermore it remains without the slightest analytical signifi- 
cance for the theories of Drucker and Mayo. 

no Recent Theories 

is 'made in USA.' . . . The true revolutionary principle is the idea 
of mass-production" (Drucker, p. i). For both authors the indus- 
trial enterprise of production is "the decisive, the representative and 
the constitutive institution" of the new order (Drucker, p. 27). Both 
have similar names for this "new order": "industrial order," "indus- 
trial society" (Drucker), "industrial civilization," "modern industrial 
society" (Mayo). However, neither for Drucker nor for Mayo are 
these categories mere sociological concepts j rather, they indicate a 
model in the sense of a desirable ideal. Thus both authors share an 
inclination to profess values and social policies which cuts through 
their analyses of reality at many points. 

Drucker's and Mayo's line of argument — neither of them likes 
the term "theory," because they identify it with practical uselessness, 
and thus we shall refrain from applying it to their conceptions — 
can be summarized in three main points. First, they begin their 
analyses with the model of an industrial society. Cooperation of in- 
dividuals and groups is the supreme principle of such a society. It is 
"a general principle for organizing people to work together" (Druck- 
er, p. 3), "a balanced relation between various parts of the organi- 
zation, so that the avowed purpose for which the whole exists may 
be conveniently and continuously fulfilled" (Mayo, p. 45). The 
structure of this society does not generate any conflicts that cannot 
be completely solved j in it, "split allegiance" is converted into "twin 
allegiance" (Drucker, pp. 146 f.), and a "common interest" combines 
all (Mayo, p. 127). There still are different group interests, even 
certain conflicts — an aspect emphasized rather more by Drucker than 
by Mayo — but their elimination is merely a matter of "intelligent 
organization that takes careful account of all the group interests 
involved" (Mayo, p. 128). "The proper study of mankind is or- 
ganization" (Drucker, p. 263). 

Second, both Drucker and Mayo admit that industrial societies 
were in an early stage, before the "revolution of mass-production," 
far removed from this model. Capitalist society had marked elements 
of a closed (Mayo: "established," Drucker: "traditional") society, 
dominated by permanent (class) conflicts, "a confused struggle of 
pressure groups, power blocs" (Mayo, p. 7), and many other dis- 
turbances. A number of historical trends have already largely abol- 
ished this state of affairs. Drucker mentions the separation of owner- 
ship and control, the emergence of the "new middle class," the level- 
ing of status, the extension of citizenship rights, the institutionali- 

Recent Theories 1 1 1 

zation of class conflict. But as yet the model — an ideal for both, but 
a realizable ideal — has not been realized properly. One element is 
stilJ lacking, the absence of which at the same time explains all dis- 
turbances and conflicts of industrial societies to the present dayj and 
this element is psychological in nature. 

For, third, the central thesis which overshadows all other con- 
siderations for Mayo, but is little less important for Drucker, can 
be summarized in the statement that conflicts and tensions such as 
those which class analysis is supposed to explain constitute but a "de- 
viation" from a normal state of human attitudes and actions, and 
can and must therefore be eliminated by "education." Mayo thinks 
that "Marx detested *the bourgeoisie' on grounds that will some day 
probably be shown to have been personal" (p. 120). The same holds 
for the labor leaders whom Mayo encountered. "These men had no 
friends. . . . They had no capacity for conversation. . . . They 
regarded the world as a hostile place. ... In every instance the 
personal history was one of social privation — a childhood devoid of 
normal and happy association in work and play with other children" 
(p. 24). Thus, class conflict was but a relapse into barbarian condi- 
tions, an expression of human imperfections, and it is necessary to 
render it impossible by the formation of "social skills," i.e., the edu- 
cation of cooperative and peace-loving men. "Where cooperation is 
maintained between the individual and his group, the group and the 
union, the union and management, the personal sense of security and 
absence of discontent in the individual run high" (p. 128). Drucker, 
who, by contrast with Mayo, is not a university professor, is a little 
more careful in his formulations. But he, too, works with Mayo's 
concept of "social skills" (p. 23); for him, too, social conflict and 
its elimination is largely a matter of "managerial attitude" (pp. 
158 ff.), a problem of "communication" and mutual understanding 
(pp. 191 ff.). "The individual" must obtain "status and function 
in the industrial enterprise" (p. 165), must learn to understand its 
goals and purposes as his goals and purposes, and must be induced 
to "responsible participation as a citizen" (p. 156) for the "industrial 
order" to function properly. What Drucker calls "integration" and 
Mayo "cooperation" is based for both "on understanding and the 
will to work together rather than on force" (Mayo, p. 115). 

The "right attitude" of individuals, or "force" — an intermediate 
reality, such as social structure, does not exist at all for Mayo and 
evaporates for Drucker wherever its patterns might disturb his policy 

112 Recent Theories 

recommendations. But I shall refrain here from a comprehensive 
critique of the conceptions of Drucker and Mayo, and those who fol- 
lowed them.^^ Instead, I shall confine myself to those aspects of 
this conception the critical examination of which may advance my 
own investigation. The first question will therefore have to be. Does 
this "conception" contain a theory which at any point affects the 
problem of class conflict in industrial societies? The answer to this 
question depends on how strict a concept of theory we employ. 

Expressed in a formula, it is Drucker's and Mayo's thesis that 
the class conflict of capitalist society was an (almost psychological) 
phenomenon of "deviance" from a normal state of integration and 
cooperation. Post-capitalist society tends toward this "normal state," 
although a number of educational measures are still required to bring 
it about. This thesis barely conceals a value judgment j it is not really 
a hypothesis that permits of empirical test, but a philosophical ob- 
servation about the immanent goal of social history or, more likely, 
an expression of certain political aims and desires. In either case 
I should hesitate to call this conception a theory. But if not a theory, 
something else is implied by Drucker's and Mayors conception which 
justifies its discussion here. It is based on a notion of society which 
Mayo makes explicit when he says, "A society is a cooperative sys- 
tem" (p. 115). The two essential ingredients of this notion are the 
assumptions that social conflict is not an essential and necessary feature 
of social structure, and that the variables which explain conflict, and 
consequently "order" and "integration," are psychological in nature. 
This basic theoretical attitude, from which a variety of analytical 
consequences follow, is by no means confined to Drucker and Mayo. 
In fact it dominates, if in sometimes rather more complex and subtle 
forms, much of contemporary American sociology, including the work 
of its most eminent theorist, Talcott Parsons. If, however, a basic 
attitude of this type proves sensible, then there is indeed no point in 
a theory of class conflict, and we should have to search for new tools 
of analysis. 

It is difficult to examine "basic attitudes" of scientific analysis 
with respect to their usefulness. The question of empirical rightness 
or wrongness does not apply to them. That "society is a cooperative 
system" is a statement which can neither be confirmed nor refuted 
by empirical propositions. We are dealing here with "meta-theoreti- 

^^ Especially in Mayo's case, this critique has been carried out several times, for 
which see as the most recent, and best, example the essay by H. L. Sheppard (87). 

Recent Theories 113 

cal"^^ decisions which determine the direction of analysis with respect 
to specific problems without being part of this analysis themselves. 
Their test is their analytical fruitfulness and not their empirical cor- 
rectness or logical soundness. We have to ask whether an image of 
society as an integrated system in which destructive conflicts occur 
only as deviations of a psychological nature from a normal state of 
order can be an appropriate background for the analysis of sociologi- 
cal problems. It is one of the themes of this study to reject and super- 
sede this image of society, and the works of Drucker and Mayo pro- 
vide a welcome opportunity to substantiate the point of view which 
underlies our considerations. 

If it is true that we have to regard society as an integrated "co- 
operative system," and that deviations from this integration must 
be explained in terms of psychological variables, it would follow, for 
example, that all socialists in a capitalist society are in some sense 
psychologically deficient, that they are "deviants." Mayo has seen 
this extreme implication; he tries therefore to explain the work of 
Marx as well as the actions of the labor leaders he met by reference 
to their "case history." There is little doubt that some recent studies 
of social psychologists with their attempt of a correlation between 
political attitudes and personality types" have lent themselves to the 
support of this conclusion. However, the hypothesis seems reasonable 
and has not so far been refuted, that there is no significant corre- 
lation between, say, a voting decision for a conservative or progressive 
party and neurotic dispositions. Moreover, even if there was such a 
correlation, one would have to find out whether its causes are purely 
a matter of individual history or of social conditions. For this is the 
image of society which we want to oppose to that of Drucker and 
Mayo: that societies create out of their structure with predictable 
certainty the conditions of social antagonisms, and that therefore 
society is not an integrated cooperative system but at best a relatively 
integrated system of conflicting structural forces, even more, a per- 
manently changing structure of integrative and disruptive factors. 

^^ An expression which in its specific meaning intended here I have taken from 
an unpublished essay of U. Torgersen (Oslo), and which indeed seems a plausible 
description of attitudes that guide empirical research without themselves permitting 
of empirical test. 

-^ I am referring here, e.g., to research on the subject of "authoritarian person- 
alities" and on national, political, and ethnic stereotypes. It would seem to be an 
important task of scientific criticism to examine such studies from the point of view 
here advanced and to ask in what sense their very design excludes the assumption of 
conflicts generated by social structure. 

114 Recent Theories 

Another extreme implication of the integration image of society 
is that "social problems" can in principle be solved only by influ- 
encing, "improving," and "normalizing" individuals. Elton Mayo 
has realized this implication, too, both in theory and in practice, in 
the context of the Hawthorne experiment in the I920^s and early 
1930's. But it seems to me that H. L. Sheppard was right in accusing 
Mayo of systematically underestimating "economic and political de- 
terminants" and problems of the distribution of power, and of re- 
ducing all conflicts to "person-to-person relations" (87, p. 327). In 
so far as refutation is possible here, it appears sufficient to recall the 
causes and consequences of numerous institutional changes of recent 
social development in the spheres of economy, state, education, etc. 

It is neither possible nor perhaps desirable to arrive at a final 
decision about which image of society is "better" or "more correct." 
But examples such as those mentioned here suggest that we reject 
the integration model as insufficient and deal with problems of con- 
flict on the basis of a different, more appropriate model. According 
to this new model, conflict is an essential element of the structure of 
every society. It grows out of this structure and can be eliminated 
only very temporarily, and only by structural changes. The carriers 
of conflict are of course individuals, but these only in so far as the 
impact of their action is directed by structural conditions to larger 
aggregates of individuals. Psychological factors are a secondary 
characteristic, not the cause of social conflict. Society is process; its 
order lies solely in the lawfulness of its change. 

It would not have been difficult to expose the insufficiencies and 
defects of Drucker's and Mayors analyses on the level of empirical 
criticism. But the deepest shortcoming of their works and those of 
others becomes apparent only if we examine their meta-theoretical 
conception of society. If Drucker and Mayo are right, then not only 
are there no classes, but there never have been such systematic con- 
flict groups in history. The phenomenon of structural conflict itself 
loses its reality and, indeed, its potentiality. However, the theorists 
of integration would find it hard to hold on to this assertion in their 
analyses. There are too many problems for which it fails to account, 
and we shall therefore be well advised to operate with a more dynamic 
image of society. 


None of the theoretical approaches of modern sociology which 
we have examined in this chapter appears to provide an entirely satis- 

Recent Theories 115 

factory solution of our problem. Although every one of them tries 
to incorporate one or another element of the new reality of an ad- 
vanced industrial society as it emerged since Marx's time, although 
therefore all of them go beyond the Marxian theory of class, none 
of them succeeds in superseding Marx's theory by a new and simi- 
larly comprehensive formulation. Four main reasons for this failure 
have emerged. ( i ) Some sociologists confine themselves to demon- 
strating that Marx's predictions have not come true and that, there- 
fore, his theory has not been confirmed. If they indicate certain "new 
lines," as Geiger does, these remain unconnected and merely confirm 
the uselessness of Marx's theory without replacing it by a new ap- 
proach. (2) The trend of sociological analysis founded on meta- 
theoretical assumptions that deny the possibility of analyzing social 
conflict as a structural phenomenon intrinsically rejects, of course, the 
very idea of a theory to replace Marx's. Thus, for Drucker and 
Mayo there is no point in any theory of conflict, since systematic 
antagonisms have no place in their image of society. (3) Some promi- 
nent sociological theories of post-capitalist society remove the subject 
matter of analysis to aspects of social structure other than those dealt 
with by Marx in his class theory. For this reason T. H. Marshall's 
and H. Schelsky's contributions remain marginal to our problem and 
cannot be considered as superseding the old theory. (4) Finally, 
there are some theoretical approaches which retain the heuristic in- 
tention of Marx's theory but fail to account for more than one or 
two changes that have occurred since. Thus Burnham's thesis is, 
contrary to the claim of its author, in fact confined to the realm of 
industrial production and ignores problems of political structure as 
well as changes in the skill structure of labor, the institutionalization 
of class conflict, etc. Djilas restricts himself to the analysis of con- 
temporary Communist societies, and his model cannot easily be 
applied to other countries. Renner's approach is, even apart from its 
untenable conclusion, merely an indication that wants elaboration 
and completion. 

Without doubt, every one of the theories discussed in this chapter 
contributes something to our knowledge of contemporary society and, 
more particularly, to our understanding of social conflict in post-capi- 
talist societies. But with respect to the precise formulation of a theory 
of social class or its equivalent, we can learn more from their weak- 
nesses and mistakes than from their substance. Our problem is the 
explanation of systematic social conflicts in industrial societies. In 
order to solve this problem, we shall have to find a formulation that 

1 1 6 Recent Theories 

passes beyond the mere statement of facts. It appears advisable to 
base this formulation on an image of society that permits the expla- 
nation of conflicts in terms of structural, not individual, conditions. 
For the theory to be useful, it will be necessary to define with utmost 
precision the area of problems for which it holds and for which it 
does not hold. Finally, the theory will have to be capable of account- 
ing for the society with which Marx was concerned as well as for 
contemporary society and for the changes that have transformed the 
former into the latter. 


A Sociological Critique of Marx 


"The relation of succeeding generations to the phenomenon of 
class society has been determined," or so Geiger thinks, "to the present 
day by the doctrine of Marx" (46, p. 10). Lipset and Bendix are 
of the opposite opinion: "The study of social classes has suffered in 
the past from the proclivity of social scientists to react against the 
influence of Karl Marx" (55, p. 151). There is probably an element 
of truth in both of these statements. Only too long has discussion 
in social science been dominated by the attempt either to reject the 
doctrine of Marx altogether or to sustain it without qualification/ 
Implicitly, if not explicitly, this attempt underlies the endless dis- 
putes about "what Marx really meant" (as G. D. H. Cole called one 
of his books). It is not difficult to see why this has happened. There 
is, for one thing, the political attractiveness, or repulsiveness, as the 
case may be, of Marx's work j there is, secondly, the prophetical prom- 
ise of his predictions 5 and there is, above all, what Schumpeter called 
the "imposing synthesis" of Marx's doctrine. "Our time revolts 
against the inexorable necessity of specialization and therefore cries 
out for synthesis, nowhere more loudly than in the social sciences, in 
which the nonprofessional element counts for so much. But," Schum- 
peter adds with equal right at this point, "Marx's system illustrates 
well that, though synthesis may mean new light, it also means new 
fetters" (73, p. 45). Schumpeter does not hesitate to free himself 
and his discipline, economics, from these fetters at the expense of 
the "imposing synthesis." In this we shall have to follow him for 

To ignore Marx is convenient, but it is also naive and irrespon- 

^ And the two statements quoted bear witness not only to opposing attitudes to 
be found among sociologists in general, but in particular to the differences in approach 
between European and American sociologists. To the present day the reaction against 
Marx (often coupled with complete ignorance of his work) is as widespread in the 
United States as the uncritical acceptance of Marx's theories in Europe. 

1 1 8 Sociological Critique of Marx 

sible. No physicist — if this analogy be pardoned — would ignore 
Einstein because he does not approve of his political attitude or of 
some aspects of his theories. To accept Marx toto coelo may testify 
to an honorable faithfulness but is scientifically fruitless and danger- 
ous. No physicist would abstain from attacking Einstein just because 
he happened to like the man and his work as a whole. We have started 
our investigation with an examination of the work of Marx, because 
his formulation of class theory is both the first and, as we know now, 
the only one of its kind. Today this theory is refuted, but it has not 
been superseded. Now we have to draw from Marx what is still use- 
ful, or, more precisely, we have to separate the problem of class theory 
from Marx^s class theory itself. We are concerned with the theory of 
class as a sociological instrument j from this point of view, the Marx- 
ian theory of class is in principle a matter of indifference, i.e., it inter- 
ests us only as a historical background or an object of critique. 

At an early point, we have distinguished Marx's "philosophy" 
from his "sociology." In doing this we have not meant, of course, to 
sustain Marx's "sociology" in its entirety. In an empirical discipline, 
that kind of intention can only lead to disaster. We shall now go about 
the business of dissecting Marx's "sociology," of sustaining what is 
useful in his approach and rejecting what is useless. There is no place 
for sentimental regard or even undue respect in a critical process of this 
kind. If, for example, the exclusive connection of the concept of class 
with economic conditions or structures ("relations of production") 
should turn out to be a "fetter," a useless assumption, it must be re- 
jected, no matter what Marx said, meant, or wanted. If, conversely, 
certain elements of Marx's theory of class formation stand up to the 
test of empirical evidence, they must be sustained j but, again, it is es- 
sentially a matter of indifference that it was Marx who formulated 
these elements. 

The charge of eclecticism might be leveled against a procedure of 
this kind. If this is so, the notion of eclecticism is used with some jus- 
tice, but it does not constitute a charge. Eclecticism may be considered 
a sin in philosophy, but science is essentially eclectic. In fact, a scientist 
who is not as such an eclectic is no scientist or at least a bad one. The 
unqualified acceptance of a "doctrine" — dogmatism — is the cardinal 
sin of science. 

The unbiased reader may grow weary of so many words about so 
obvious a matter, but unfortunately they are, in the case of Marx, still 
necessary. There still are many who do not see that the epithets 
"Marxist" or "anti-Marxist" have no meaning and place in a science. 

Sociological Critique of Marx 119 

In fact, this is not the least reason why the vast literature about the 
subject of class (or at least about the word "class'*) has brought so 
little real advance. It appears as if sociology has hardly progressed 
at all in this respect since Marx — surely an astonishing fact in view of 
its extraordinary development in other fields! 

However, it would be false to say that the attempt to reformulate 
the theory of class is entirely lacking in precedents and points of con- 
tact in recent sociology. We shall have opportunity to refer to numer- 
ous connections and suggestions in the course of our considerations. 
Without injuring the originality and individuality of the scholars con- 
cerned, one might even construct a certain convergence of the concep- 
tual and theoretical discussion of class in recent sociology. The works 
of Schumpeter, Renner, Geiger, Lipset, Bendix, and even Parsons 
mark some of the milestones on this road towards a sociological theory 
of conflict in industrial societies. Other contributions will be added to 
this list as we go along. Thus the present investigation does not claim 
particular originality (which in any case is a dangerous virtue in 
scholarship). It is rather an attempt to connect many loose threads 
into a net with which we can "catch" an important sector of social 

In the present chapter we shall be concerned with elucidating the 
prerequisites of the fruitful use of something like a concept and theory 
of class in sociological analysis. Although we shall in every case build 
our discussion around a critical examination of Marx's position, our 
assumptions will increasingly depart from those of Marx and his 
theory of class. Having ascertained the conditions and main categories 
of a class theory of social conflict, we can then proceed (in the follow- 
ing two chapters) to an attempt to outline the main features of this 
theory. In the last two chapters of this study the analytical usefulness 
of the reformulated theory of class conflict will be subjected to a test 
by its application to the structure of post-capitalist societies. 


In the last decades, considerable progress has been made in the 
development of a theoretical instrumentarium for analyzing the struc- 
ture of total societies and their parts. The credit for codifying the ele- 
ments of this analysis is principally due to American sociologists, al- 
though their most eminent representative, Talcott Parsons, has rightly 
referred (see 216) to earlier impulses in the work of the Englishman 
Alfred Marshall, the Frenchman Emile Durkheim, the Italian Vil- 
fredo Pareto, and the German Max Weber. It is too early yet to speak 

120 Sociological Critique of Marx 

of a complete theory that is at our disposal in structural analysis j all 
we have is a theoretical instrumentarium. The notions of a "func- 
tional" or "structural-functional theory" are in many ways premature, 
if one does not want to extend the term "theory" beyond its strict 
meaning. These notions refer to what is above all a set of intercon- 
nected categories (partly linked by generalizing assumptions), whose 
application to empirical problems permits the general description of 
social structures as well as the determination of the place of specific 
elements in them. Societies and organized units within societies 
(groups, associations, institutions)^ have a structure or can be regarded 
as units displaying a structure. "To exhibit the structure of an object 
is to mention its parts and the ways in which they are interrelated. . . . 
Every account of structure is relative to certain units which are, for the 
time being, treated as if they were devoid of structure, but it must 
never be assumed that these units will not, in another context, have a 
structure which it is important to recognize" (Russell, 222, pp. 267, 
269). The basic unit of structural analysis in sociology to which 
this statement obviously applies, is that of role, i.e., of a complex of 
behavior expectations which are associated with a given social position 
or status. In structural analysis, the human individual in the fullness 
of his expressions figures only as an incumbent of such positions, and 
"player" of roles. The relations between roles and their agglom- 
erations around certain institutional spheres (occupation, education, 
family, politics, etc.) are expressed by the concept of function, i.e., by 
their latent or manifest consequences for the "functioning" of the 
total structure. Thus, the structure of a society presents itself in its 
most formal aspect as a functional system the units of which are social 
roles and role sets.^ 

But by contrast with the structure of other objects of knowledge, 
especially of organisms with which they are frequently compared, so- 
cial structures have one important peculiarity. They are not as such 
"given," they cannot in principle be analyzed independent of their 

^ "Structural-functional" theorists like to refer to both with the term "social 
system." Reasonable as this link by a general category is, it is nevertheless dangerous 
because of the possible implications of the concept of system from the point of view 
of conflict and change: systems might appear as closed units which do not permit of 
change. This is, to be sure, a misunderstanding; but in order to forestall it I would 
propose to avoid the term "system." 

^ I neither intend nor can attempt here to give even the barest outline of the 
categories and assumptions of the "structural-functional" approach. For that, see the 
works of Parsons, Merton, Levy, and others; see also my essay on Parsons (206). Here 
we are concerned merely with certain formal aspects of this approach, in so far as they 
are relevant for the theory of class. 

Sociological Critique of Marx 1 2 1 

historical context, but they are themselves subject to continuous 
change. By change in this sense we do not mean the occurrence of cer- 
tain processes within a given structural pattern, for this is accounted 
for by the category of structure in any case. Regular processes within 
objects that have a structure — such as the processes of role allocation, 
or of the socialization of new members of society — are indeed an es- 
sential element of every structure j structural analysis is essentially the 
analysis of such processes. What is meant here is, rather, that the en- 
tire structural arrangement of so-called forms of society can change. 
Function and functional importance of the heart or liver in an organ- 
ism do not change ; function and functional importance of religious or 
economic institutions in society not only can change but also are subject 
to a continuous process of change in all known societies. The function- 
ing of medieval European society would evidently be strongly en- 
dangered if in an experiment of imagination we removed all religious 
institutions from this society ; were we, however, to remove these insti- 
tutions from today's secularized industrial societies the effect would 
be considerably milder. An analogous argument could not be made 
for organisms. Should we remove the heart from the human organ- 
ism, the consequences would be the same at all times. Radcliffe-Brown 
has realized this peculiarity of social structures more clearly than 
many of the later "structural-functionalists" when he states, "that an 
animal organism does not, in the course of its life, change its structural 
type. A pig does not become a hippopotamus. . . . On the other 
hand a society in the course of its history can and does change its struc- 
tural type without any breach of continuity" (220, p. 181). 

Russell remarks in his logical analysis of the concept of structure, 
"An analysis of structure, however complete, does not tell you all 
that you may wish to know about an object. It tells you only what are 
the parts of the object and how they are related to each other 5 it tells 
you nothing about the relations of the object to objects that are not 
parts or components of it" (222, p. 268). In the case of organic struc- 
tures, this involves a limitation of structural analysis, not an objection 
against it. Anatomy and physiology have heuristic value and scientific 
validity even without a social psychology of relations between organ- 
isms. Social structures, however, carry within them the seed of other 
structures that lie beyond their (fictitious) borderlines. They reach, 
so to speak, beyond themselves j at any given point of time they either 
are no longer or not yet what they appear to be. Process and change 
are their very nature and indicate therefore superordinate categories 
of analysis. Although in biology the analysis of the evolutionary 
process can rest on structural analysis, such analysis must, in sociology. 

122 Sociological Critique of Marx 

be subordinated to the analysis of processes of change of structural 

Time and again, structural-functional theory has been accused of 
not recognizing this basic fact of social reality. With respect to the 
intention of most of its advocates at least, this accusation is unjustified. 
Not merely, but especially, in sociology the analysis of changes of 
structural patterns creates almost insuperable problems. "We reason 
about movement," so Bergson laments on behalf of philosophy, "as 
if it were composed of immobilities, and if we consider it, we compose 
it of immobilities. Movement for us is a position" — such as a struc- 
ture — "and then a new position, etc., ad infinitum^* (223, p. 165). 
The statement is justified, but not the lament. It appears, rather, as if 
processes are accessible to our analysis only, if we dissect them into 
their static elements 3 more precisely, if we try to reconstruct them 
from a static basis (which changes) and from certain forces (which 
cause change).* Talcott Parsons has been acutely aware of this prob- 
lem of knowledge. For him, the concept of structure is therefore no 
more than an inevitable expedient, no more than this constructed 
"static basis": "Structure does not refer to any ontological stability in 
phenomena but only to a relative stability — to sufficiently stable uni- 
formities in the results of underlying processes so that their constancy 
within certain limits is a workable pragmatic assumption" (217, 
p. 217). Parsons and many other sociologists have also seen that the 
next step of analysis must lie in the designation of the dynamic ele- 
ments of social structures. But in carrying out this designation they 
make that central mistake which renders a large part of their cate- 
gories useless for the analysis of structural change, and which there- 
fore justifies the charge of a "static bias" in their approach. 

Parsons continues the statement quoted above by saying: "Once 
resort is made to the structure of a system as a positive constituent of 
dynamic analysis there must be a way of linking these ^static' structural 
categories ... to the dynamically variable elements in the system. 
This link is supplied by the all-important concept of function. Its 
crucial role is to provide criteria of the importance of dynamic factors 
and processes within the system. They are important in so far as they 
have functional significance to the system." The consequential mis- 

* There is another way of looking at the same logical situation, and one that in 
the long run may prove more appropriate. In a Galilean sense, one might assume move- 
ment as the normal state of affairs, and instead of looking for forces that inaugurate 
change one might concentrate on forces which arrest movement or slow it down. In 
Chapter VI below we shall explore the applicability of this kind of approach to prob- 
lems of social change. 

Sociological Critique of Marx 123 

take of this step of analysis already lies in its intention to find the dy- 
namically variable elements "in the system." The category of func- 
tion is indeed subordinated to that of structure. Parts of a structure 
have a function in relation to this structure as a whole. In this sense 
the category is certainly important. However, it is not "all-impor- 
tant": it is rather the first requisite of a dynamic analysis of structure 
to find variables which are not subordinated to the category of struc- 
ture (and are in this sense "within the system") but which operate as 
forces or factors changing the structure. That Parsons, and with him 
many other recent "theorists," have overlooked this fact may be due 
to a more or less deliberate identification of organic and social struc- 
tures or "systems." For this is the most difficult problem of the analy- 
sis of structural change: by contrast to organic structures, the "dynami- 
cally variable elements" which influence the construction of social 
structures do not necessarily originate outside the "system" but may 
be generated by the structure itself. There are, in other words, within 
social structures certain elements or forces which are at the same time 
their constituent parts (and therefore "function" within them) and 
impulses operating toward their supersedence and change. As we shall 
see, social classes are elements of this kind. 

It is neither necessary nor possible here to examine the implica- 
tions of this critique in detail. Moreover, we shall have to confine 
ourselves to these few and necessarily abstract remarks, the empirical 
relevance of which will emerge more clearly when we turn to the dis- 
cussion of class or role interests. Structural-functional analysis as it 
stands today fails to explain problems of change because it does not 
account for the peculiar character of social as opposed to organic struc- 
tures. It does not look for the dynamic variables that, though oper- 
ating within given structures, are in principle independent of their 
(constructed) functional integration. If, as is the undisputed premise 
of all sociological inquiry, we are ultimately concerned with the scien- 
tific description and explanation of structural change, then we must 
find, apart from the undoubtedly important construction of function- 
ally integrated structures, elements which are independent of these 
without being necessarily external to them and which determine rela- 
tive stability as well as kind and degree of change of structural pat- 
terns. In identifying such factors we shall have to be careful not to 
abandon the gain in systematic analysis achieved by the structural- 
functional approach to the randomness of factors adduced ad hoc. The 
emergence and operation of forces that change social structure are also 
subject to laws which we may want to recognize. 

Even the careful reader of Marx may ask himself, in vain, in what 

1 24 Sociological Critique of Marx 

sense the rather abstract discussion of this section sustains Marx's posi- 
tion (as the heading of it promises). Indeed, this discussion shows 
how far sociology has advanced since Marx. In the place of undiffer- 
entiated and often implicit premises and assumptions we have today 
almost too elaborate categories and theoretical standpoints. Yet im- 
plicit and perhaps not even conceived with full clarity as it may be, 
the right approach to dynamic social analysis can be discerned in the 
works of Marx. Throughout his works, Marx displayed a strong con- 
viction of the primacy of the analysis of structural change. He, too, 
had to construct the model of a society ("capitalism") for this pur- 
pose. But he did not stop there. Being intensely concerned with dis- 
covering pattern in history, he searched for factors and forces that 
promise to explain the process of social change. As we shall presently 
see, Marx overstated his case, to say nothing of making numerous 
mistakes in details of his theory. But in so far as historical societies 
are concerned, Marx never fell into the trap of abandoning the prob- 
lem of change out of fascination with the beauty of his structural 
model. His subject was social change, and the category of social struc- 
ture was no more than a tool with which to tackle this elusive and 
intricate problem. 


Many aspects of Marx's theory of class have to be rejected in the 
light of sociological knowledge. These do not include, however, the 
heuristic purpose of Marx's sociological work and its immediate con- 
sequences. It is without doubt important to develop categories to de- 
scribe social changes. Concepts like "role differentiation," "transfer- 
ence of functions," "leveling of statuses," and the like serve this pur- 
pose. But it is clearly more important to find ways and means to ex- 
plain change. It is of course most unlikely that any one hypothesis 
will be capable of accounting for all types of change that can be ob- 
served in the course of history, and in so far as Marx advances an abso- 
lutist claim for his own theory we shall have to depart from him radi- 
cally. At the same time, Marx has explored one of the most interest- 
ing, and perhaps the most significant, relationship between social struc- 
ture and social change by postulating conflict groups and their clashes 
as forces that make for change. Obvious as it may seem that social 
conflicts often result in the modification of accepted patterns of organi- 
zation and behavior, it has neither been seen by all nor been explored 
as systematically by anybody as by Marx. 

Throughout his life, Marx was clearly influenced by the memory 

Sociological Critique of Marx 125 

of two events which overshadowed the consciousness of the nineteenth 
century, although they were its heritage rather than its product: the 
French Revolution and the Industrial Revolution. There were obvi- 
ous explanations for both of them, and these were widely held. One 
might summarize them by the phrases that "men make history" and 
that "inventions make history." Even today, historians find it hard 
to free themselves from the conception that at the turning points of 
history there stood outstanding and powerful individuals or important 
and consequential inventions. Indeed, it would be nonsensical to try 
and deny the effect of these forces. But, as Marx well knew, in the 
French Revolution and the Industrial Revolution another kind of 
force had also become apparent. Apart from powerful individuals 
and revolutionary inventions, larger and more anonymous aggrega- 
tions of men had played a visible part in bringing about these events. 
Nor had they been unanimous in purpose and action. It was, rather, 
the conflict between aggregates of differing (but considerable) size, 
and the changing fortunes of this conflict, that had effected a restruc- 
turing of society so far-reaching that it could only be called revolu- 
tionary. As we shall see, the revolutionary tradition of the eighteenth 
century not only inspired Marx but misled him as well. He tended to 
believe that the only way in which social conflicts could produce struc- 
tural changes was by revolutionary upheavals. But despite such errors 
he did discover the formative force of conflicting social groups or 
classes. This "discovery"^ is accompanied, in the work of Marx, by 
two steps of analysis which, although rather formal, are nevertheless 
worth mentioning and sustaining. 

Firstly, Marx succeeded in tracing conflicts that effect change back 
to patterns of social structure. For him, social conflicts were not ran- 
dom occurrences which forbid explanation and therefore prediction. 
Rather, he believed these conflicts to be necessary outgrowths of the 
structure of any given society and, in particular, of capitalist society. 
It is doubtful whether Marx, by assuming property relations to be the 
structural origin of conflict, was right in the substance of his analysis. 
But this does not diminish the analytical achievement of tracing in the 
structure of a given society the seeds of its supersedure. The idea of 
a society which produces in its structure the antagonisms that lead to 

" Like all discoveries, it is not, strictly speaking, original. It would not be diffi- 
cult to find, throughout the history of pre-Marxian philosophy from Heraclitus to 
Hegel, numerous thinkers who regarded "conflict as the father of all things." But it 
is Marx's merit to have embodied this approach in a fairly systematic theory of the 
generation and course of social conflict in modern societies. 

126 Sociological Critique of Marx 

Its modification appears an appropriate model for the analysis of 
change in general.^ 

Secondly, Marx properly assumed the dominance of one particu- 
lar conflict In any given situation. Whatever criticism may be required 
of the Marxian theory, any theory of conflict has to operate with some- 
thing like a two-class model. There are but two contending parties — 
this is implied in the very concept of conflict. There may be coalitions, 
of course, as there may be conflicts internal to either of the contenders, 
and there may be groups that are not drawn into a given dispute j but 
from the point of view of a given clash of interests, there are never 
more than two positions that struggle for domination. We can follow 
Marx in this argument (which, for him, is often more implicit than 
explicit) even further. If social conflicts effect change, and if they are 
generated by social structure, then it is reasonable to assume that of the 
two interests involved in any one conflict, one wiU be pressing for 
change, the other one for the status quo. This assumption, again, is 
based on logic as much as on empirical observation. In every conflict, 
one party attacks and another defends. The defending party wants 
to retain and secure its position, while the attacking party has to fight 
it in order to improve its own condition. Once again, it is clear that 
these statements remain on a high level of formality. They imply no 
reference to the substance or the origin of conflicting interests. But, 
again, it will prove useful to have articulated the formal prerequisites 
of Marx's and, indeed, of any theory of conflict. 

With these formal points, however, our agreement with Marx 
ends. Although the heuristic purpose and general approach of his 
theory of class can and must be sustained, this is not the case with 
respect to most other features of this theory. Only by rejecting these 
can we hope to clear the way for a more useful theory of class conflict 
in industrial societies. 


Since Talcott Parsons wrote his Structure of Social Action the neg- 
lect of a systematic analysis of the dynamics of social action by sociolo- 
gists has become increasingly conspicuous. Only very recently have a 
number of scholars set out to explore and map this white spot in the 

® Parsons justly emphasizes, in an essay on Marx that is surprising in more than 
one sense, this achievement of Marx and states that Marx "did . . . unlike the utili- 
tarians, see and emphasize the massive fact of the structuring of interests rather than 
treating them at random" (67, p. 323). 

Sociological Critique of Marx 127 

atlas of sociological knowledge. If only for this reason, it is of some 
importance to determine the logical status and limits of dynamic analy- 
sis rather more precisely than is necessary today with respect to prob- 
lems of, say, social stratification. We have tried to reduce the spongy 
concept of social change to that of structural change. This constitutes a 
gain, but it is not in itself sufficient. At a later point we shall have to 
return to the dangerous question "When does a structure begin to 
change or, conversely, up to what point does it remain unchanged? " — 
a dangerous question because it implies an essentially static concept of 
structure. So far we have merely touched upon the two cardinal re- 
quirements of a theory of change, i.e., the construction of the model of 
a functionally integrated structure, and the discovery of certain factors 
or forces the effect of which leads to a modification of this structural 
model. As to the first of these requirements we have, in the structural- 
functional approach, a considerable instrumentarium at our disposal 
today. But with respect to the codification of forces that effect struc- 
tural change everything is still to be done. Ad hoc and at random fac- 
tors are introduced wherever necessary, and all too often these factors 
are afterwards generalized in an impermissible manner. Thus we get 
so-called theories of the primacy of the economy, of race, of elites, of 
cultural diffusion — or of classes. We cannot hope to remedy the obvi- 
ous lack of a systematic treatment of this subject by preliminary classi- 
fications and delimitations as will be proposed here j but we can try to 
avoid the most obvious errors of one-sided theories. In order to do so, 
however, it will be inevitable that — as T. H. Marshall says with 
pleasant irony of his own investigation (57, p. 10) — "I shall be run- 
ning true to type as a sociologist" by proposing to divide our subject 
into several distinct parts. 

Among the forces that are capable of changing elements of social 
structure, two large groups must evidently be distinguished — those 
that originate outside a given structure and those that are generated by 
the structure itself. We shall use for the former the concept of exoge- 
nous structure change, or exogenous factors, and for the latter that of 
endogenous structure change, or endogenous factors.^ 

^ M. J. Levy has introduced a similar distinction (209, p. 114). "The strategic 
factors for change (i.e., the factors necessary and sufficient for a change given the 
initial stage) may be internal factors (i.e., factors produced by the operation of the 
unit without any new influences from other units), or external factors (i.e., factors 
newly introduced to the system from other units), or some combination of the two." 
Unlike us, however. Levy is above all concerned "with cases of strategic external 

128 Sociological Critique of Marx 

If the invasion of an African territory by European conquerors 
causes the abolition or modification of the chieftain system in certain 
tribes, we are faced with exogenous structure change. But the separa- 
tion of ownership and control or the institutionalization of class con- 
flict in post-capitalist society are endogenous, whichever factors one 
may identify as responsible for these changes. It is clear that this 
distinction is strictly possible only in analytical and not in empirical 
contexts. In conspicuous structure changes in particular, such as the 
industrial revolution, exogenous and endogenous forces usually com- 
bine to produce the change. It is an important task of the empirical 
analysis of specific problems to disentangle the two and assess their 
respective weights. 

Within each of these fields of factors further distinctions are re- 
quired. Thus, exogenous change can result from military conquest 
and deliberate intervention with existing structures 5 but it can also 
result from the diffusion of culture patterns unaccompanied by politi- 
cal or military force. In past decades, many efforts have been made, 
above all by social anthropologists, to bring the different forms of 
exogenous change into an ordered context by introducing such concepts 
as "diffusion," later "acculturation," "culture contact," and "culture 
change." Empirical instances of some of these forms have been stud- 
ied in great detail. But despite Malinowski's attempt to systematize 
such approaches (213), this effort has not yet advanced beyond a loose 
catalogue of possible factors. The sociology of war and of contacts 
between advanced societies (until now an utterly neglected field of 
study) could also contribute to closing this gap in theory. 

However, we are not only no further, but possibly less advanced, 
with respect to the classification of forces operative in endogenous 
structure change, although many seem to think that this is the proper 
subject matter of sociological inquiry. Although the number of such 
factors proposed by sociologists to fulfill requirements of research 
(and sometimes demands of philosophical or political convictions) 
grows steadily, a systematic examination of these factors and their 
interrelations has not even been attempted. The matter is further 
complicated by the fact that in some cases, such as the differentiation 
of roles or functions and of technological processes, we can hardly 
venture a guess as to which factors contribute to their emergence, to 
say nothing of certain knowledge about them. Marx's attempt to con- 
nect the development of productive forces with that of classes marks 
one of the weakest points in his sociology. It appears most improbable 
that the complication of the social division of labor, or technological 

Sociological Critique of Marx 129 

processes that have social consequences, can be explained in terms of 
group conflicts. In any case, structure changes resulting from social 
conflicts between organized groups or between the representatives of 
unorganized masses constitute but one form of endogenous change. 

Even within this considerably restricted sphere of social conflicts 
afi^ecting structure change it is not only possible but necessary to dis- 
tinguish a plurality of different forms. It obliterates the precision of 
analysis if, with one and the same set of categories, we try to analyze 
conflicts between slaves and freemen in ancient Rome, Negroes and 
whites in the United States, Catholics andProtestants in contemporary 
Holland, capital and labor in capitalist society — to mention only a few 
possibilities. All these conflicts can result in structure changes; they 
are in this sense factors of endogenous change. Moreover, several of 
these types of conflict may be superimposed on each other, and may 
thus constitute a single conflict front in a given country and situation. 
For purposes of analysis, however, it is necessary to introduce distinc- 
tions if one wants to master reality with the tools of science.* Endog- 
enous change is but one kind of social structure change; social con- 
flict is but one of the causes of endogenous change; and class conflict 
is but one type of social conflict. Endogenous change may be of great, 
even dominant, significance in a given society; but that is a matter for 
empirical research. In principle a theory of class illuminates only a 
small segment of the wide field which can be described by the vague 
concept of structure change. We can neither expect nor, above all, 
assume that a theory of class will cast a glimmer of its light on other 
aspects of structure change as well. 

It is apparent that from this point of view Marx is in a sense guilty 
of the same mistake of which, in a diff^erent context, we have accused 
those who have endeavored to supersede the theory of class with a 
theory of stratification. They, too, have illegitimately transposed a 
theory from its legitimate place to other areas of inquiry. The asser- 
tion that the history of all past society is the history of class struggles 
is either meaningless or false. It is meaningless if it is merely intended 
to say that, inter alia, there were also class conflicts in every society. 
But Marx did not mean this. He believed that the dominant conflicts 
of every society were class conflicts, and indeed that all social conflicts 
and all structure changes can be explained in terms of antagonisms of 
class. This generalization is as impermissible as it is untenable. The 

^ In an earlier article 1 have indicated a classification of tvpes of endogenous con- 
flicts, distinguishing "partial conflicts" (minorities), "sectional conflicts" (town- 
country), and class conflicts (41, p. 175) ; but this is only a beginning. 

130 Sociological Critique of Marx 

place to which we have assigned the theory of class may appear modest. 
In fact, this reduced significance of class analysis will be corrected to a 
certain extent with respect to the particular historical constellations of 
industrial societies. This does not alter the fact, however, that social 
analysis in terms of class — as Gurvitch remarks quite rightly (50, p. 
290) — "does not by any means provide a key which opens all doors 
to the solution of problems of social change." I would claim that only 
by restricting the theory of class to one, if one major, aspect of struc- 
tural change, can we hope to weld it into a useful tool of sociological 


"The conflict between proletariat and bourgeoisie" is, according 
to Marx, "a conflict of one class against another, a struggle which in 
its highest expression means a total revolution" ( 6, pp. 1 8 8 f . ) . More 
generally, the revolutionary character of social change is for Marx 
a central feature of his theory of class. Indeed, it appears to be Marx's 
conviction that wherever classes exist structure change has always and 
"necessarily" a revolutionary character. "Only in an order of things 
where there are no classes and no class conflicts will the social evolu- 
tions cease to be political revolutions" (6, p. 1 89). It emerges clearly 
from Marx's writings, and has never been doubted by his interpreters, 
that the concept of revolution meant for him the sudden and rapid 
upheaval of a social structure ; Marx did not speak of revolutions in 
the extended sense of a "managerial revolution" and the like.^ This 
conception, according to which social change occurs suddenly and by 
widely visible explosions, has not been confined to Marx and his faith- 
ful followers. Even Brinkmann betrays a trace of the conviction that 
social changes are always revolutionary if they result from class con- 
flict, when he observes an "evolutionary moderation of revolutionary 
forces and patterns" in recent social development (192, p. 12). Yet, 
here again we find one of those untenable generalizations which bar 
the path to our knowledge of reality and which have to be replaced by 
more reasonable, if not empirically confirmed, assumptions. The error 
that changes of social structure are generally of a revolutionary nature 
is especially interesting for two reasons which are worth pursuing. In 

^ It is however true that Lassalle and, following him, Renner have in the course 
of their revision of Marx replaced the notion of the "revolution in the hay-fork sense" 
(Lassalle) with the notion of gradual social changes. But in doing so, these men were 
aware of the fact that they had abandoned a central tenet of Marx's theory. 

Sociological Critique of Marx 131 

the first place this very assumption might, despite its dynamic appear- 
ance, induce us to join in with Bergson's lament about dissolving 
change into "immobilities." In the thesis that a given structure can be 
changed and transformed into a new one only by radical upheavals, 
there is ultimately an assumption that social structures are basically 
static entities. Of course, Marx spoke of the "law of development," 
that is, of the dynamics of capitalist society. But this law of develop- 
ment was for him little more than the law of development of an or- 
ganism: the gradual unfolding of a "system" to its inherent image. 
The structure, or "system," is as such immutable. If it changes, it is 
destroyed entirely. It changes in one stroke (or at least by one 
"stroke") in all its parts and thus becomes a new "system." This is a 
point at which Marx and Parsons meet in a curious fashion: both of 
them freeze the flow of the historical process in the idea of a "system." 
If we accept this, either structure change can be nonexistent (which, 
by an extreme interpretation, might be called the Parsonian "solu- 
tion"), or it exists only as revolutionary change (the Marxian "solu- 
tion"). Both those solutions are equally unsatisfactory and untenable. 
They testify, moreover, to the insufficiency of all conscious or uncon- 
scious analogies between organic and social "systems." 

If changes of social structure are invariably revolutionary in char- 
acter, there can be no change without revolutions. It is easy to see how 
at this point the sophistic argument offers itself that Western indus- 
trial societies have remained unchanged since Marx because they have 
not experienced revolutions. But it is also easy to see how miserably 
an assertion of this kind fails to account for the processes of reality. 
We have seen that many recent sociological theories of class conflict 
refer to a "revolution" in the social development of past decades. The 
use of the term "revolution" for processes which are neither sudden 
nor explosive bears witness to the extent to which "revolution" and 
"change" are interlaced in general opinion. Quite contrary to such 
ill-considered formulations, it is the decisive characteristic of the de- 
velopment of industrial societies since Marx that profound structure 
changes have occurred without sudden, widely visible upheavals. In 
this sense, social development of the past decades furnishes evidence 
for the ubiquity and gradualness of social change and for the untena- 
bility of the assertion that change must always be revolutionary. 

For Marx, the assertion of the revolutionary nature of social 
change has another aspect which appears no less untenable in the light 
of empirical knowledge. Before Marx, Hegel had, when analyzing 
the dialectics of "wealth" and "poverty" in his Phenomenology of 

132 Sociological Critique of Marx 

Mind^ identified the "deepest depravation," "purest inequality," and 
"absolute insignificance of the absolutely significant" with the "deep- 
est rebellion" (226, p. 368). Insignificantly changed, this idea re- 
appears when Marx says that the proletariat "is forced, by the irre- 
jectable, indefensible, absolutely commanding need — the practical 
expression of necessity — to rebel against this inhumanity" (4, p. 207). 
It is only a step from here to Marx's assertion that the class struggle 
will become more intense as the life situation of the proletariat dete- 
riorates and will culminate in a revolution when the point is reached at 
which this situation has attained its extreme. Even from the point of 
view of a sociology of revolution, this assumption can be regarded as 
disproved today. Revolutions and revolts do not occur when need and 
oppression have reached an extreme point j they occur rather once this 
extreme has been passed and the lethargy given with it superseded." 
Beyond this empirical inadequacy, Marx's notion of a linear increase 
of the violence of class conflict to the breaking-point of the revolution 
proves a Hegelian heritage which contributes little to our understand- 
ing of reality. Plausible as the application of dialectics to history may 
appear, its schematized and simplified consequences rarely survive 
empirical test. 

If in this study I refer to structure change, I do not therefore mean 
revolutions. If I refer to class conflict, I do not imply an assumption 
that it is subject to an "inevitable" process of intensification leading up 
to a revolutionary explosion. Although the construction of a struc- 
tural model is a prerequisite of the systematic analysis of change, this 
structural model must not be viewed as a monolithic entity which in 
some unknown sense can only change "as a whole." Rather, structure 
change has to be assumed as a permanent aspect of every society. It 
can begin in one sphere of a structure, such as industry, and propagate 
to other spheres, such as political society; but it can also remain con- 
fined to one sphere. Even if, for example, it could be shown that the 
separation of ownership and control in industry has no consequence for 
the political structure of society, such a separation would nevertheless 
constitute a structural change. Only if we view structure change as a 
ubiquitous and constituent element of social structure do we free our- 
selves of the fetters of the assumption that social change is always of 
a revolutionary character. At the same time we can thus avoid the in- 

^° The revolts and revolutions in the recent history of Eastern Europe (June 17, 
'953 > Poznan, Hungary) have confirmed this hypothesis quite convincingly. In gen- 
eral, these events provide many an illustration for the theses advanced in this and the 
following chapters of this study. 

Sociological Critique of Marx 133 

soluble task of determining when and where processes of change "be- 
gin" and "end." 

Regarding class conflict, one important assumption follows from 
this decision. If social change is not confined to revolutionary explo- 
sions but is a constituent element of every structure as such, it is no 
longer necessary to assert a linear development of classes and class 
conflicts toward the point of revolution. For Marx, the classes them- 
selves were within a given system something like "organisms" with 
a predetermined course of development leading to their perfection. 
It follows from this that, among other things, organized classes 
begin to affect the structure within which they have grown only in the 
moment of revolution, and that their changing force is indeed confined 
to this moment. All that happens before the revolution happens 
merely in preparation for this event j afterwards, the classes of the old 
society dissolve. Again, this simplifying assumption robs the concept 
of class of its empirical value for nonrevolutionary processes, and that 
means for the majority of social processes. Again, we have to depart 
radically from Marx. The "necessity" of a linear intensification of 
class conflict is in any case a nonsociological postulate which has to be 
rejected. Moreover, it will prove reasonable to view the interrelations 
of classes as potentially a process of permanent readjustment. Periods 
of violent conflict can be succeeded by others of relative harmony, and 
vice versa. There is no general law that determines the course of 
clashes and struggles between classes j and no revolutionary upheaval 
can be postulated as a "normal" goal and climax of class conflict. The 
course of conflicts between classes presents a problem which cannot be 
solved by arbitrarily introducing assumptions or premises but has to 
be determined on the basis of empirical investigations into particular 
historical constellations in particular societies. 


In rejecting the premise of a predetermined course of class conflict, 
we implicitly deny the validity of a further assertion that is part and 
parcel of Marx's theory of class, namely, that classes are always mani- 
festly antagonistic groups or tend toward manifest conflicts. The 
assumptions rejected so far were more or less peripheral for Marx's 
theory of class, but here we come to its core. This cannot prevent us 
from rejecting untenable assumptions without hesitation, but it makes 
caution imperative, if we do not want to run the risk of disputing the 
heuristic value of any theory of class together with the problematic 
aspects of its Marxian formulation. "Individuals," Marx says at one 

134 Sociological Critique of Marx 

place, "form a class only in so far as they are engaged in a common 
struggle against another class" (13, II, p. 59). This statement con- 
tains, if it is interpreted in the light of other statements of Marx about 
the same matter, a necessary assumption and a false empirical generali- 
zation. In the interest of an analytically useful theory of class we shall 
have to separate the two. 

The theory of class aims at a systematic analysis of one of the 
causes of the endogenous structure change of societies. It has its place 
within the wider context of the analysis of structural changes caused 
by social conflicts. From this it is evident that, however one may choose 
to define classes, they must always be regarded as groupings related to 
each other in such a way that their interplay is determined by a struc- 
turally conditioned conflict of interests. In this sense, one class alone 
is a contradictio in adiecto ; there must always be two classes. In this 
sense, also, Stalin's concept of "nonantagonistic classes" is meaning- 
less j where there are classes, there is conflict. Inasmuch as any theory 
of class is a theory of structure change by social conflict, the assump- 
tion of a conflict between classes is part of the definition of classes. 
There can thus be no reason to reject Marx's formulation on this point. 

But beyond this formal statement, Marx postulated acute and vio- 
lent conflict ("class struggle") as part of the definition of classes; and 
we cannot follow him in this step. That class conflict invariably as- 
sumes violent forms and becomes civil war is an assumption the em- 
pirical character of which Marx could not abolish by joining it to his 
definition of the concept of class. "Exactly how serious the element of 
conflict is becomes a matter of empirical investigation" (Parsons, 67, 
p. 324). What evidence we have permits at least the negative con- 
clusion that class conflict does not always assume the form of civil war. 
Instead, a phenomenon like the institutionalization of class conflict 
shows that an "oppressed" class may well be capable of effecting struc- 
ture changes by discussion and negotiation. Here, as elsewhere, more 
acute analysis reveals that the overly simple assumptions of Marx tend 
to obliterate rather than illuminate the intricacies of the problem of 

Parsons introduced for purposes of a subtler analysis of class rela- 
tions the useful concept of "potential" or "latent conflicts" (67, p. 
329). Apart from the actual or manifest clashes between classes it 
seems reasonable to distinguish at least two kinds of latent conflicts. 
Marx has himself dealt with one of these, namely, with what one 
might call the immature conflicts between classes which are still in the 
process of formation and organization. But a second form of latent 

Sociological Critique of Marx 135 

class conflict seems even more important. It appears that conflicting 
classes can, for several reasons, co-exist for shorter or longer periods 
of time in a kind of "armistice" without engaging in open struggles. 
Some of the symptoms of this reduction of manifest to latent conflicts 
are well known and confirmed by considerable evidence: common 
interests, such as national interests in emergency situations, can be 
superimposed on group antagonisms for certain (limited) periods j 
conflicts can be formalized to the extent of being transformed into dis- 
cussions between plenipotentiaries or representatives in parliaments or 
industrial negotiation bodies. At this point, too, sociological analysis 
has to turn away from the sterile magic of definitional premises to the 
investigation of empirical conditions under which latent conflicts be- 
come manifest or manifest conflicts fade into the background. 

Some authors prefer to describe antagonisms and tensions which 
are not expressed in manifest struggles in terms other than conflict. 
Thus, they distinguish conflicts and tensions, conflicts and disputes, 
conflicts and contests, or — most frequently — conflict and competition. 
Such terminological distinctions are in fact in keeping with common 
usage. We do, indeed, tend to associate with the word "conflict" visi- 
ble clashes between forces, i.e., antagonisms which are manifest as 
such. A football game, a competition between applicants for a job, a 
parliamentary debate, or a legal contest are not usually called con- 
flicts. However, it will be evident from the preceding discussion that 
I am using the term "conflict" in this study for contests, competitions, 
disputes, and tensions as well as for manifest clashes between social 
forces. All relations between sets of individuals that involve an in- 
compatible difference of objective — i.e., in its most general form, a 
desire on the part of both contestants to attain what is available only 
to one, or only in part — are, in this sense, relations of social conflict. 
The general concept of conflict does not as such imply any judgment 
as to the intensity or violence of relations caused by differences of ob- 
jective. Conflict may assume the form of civil war, or of parliamen- 
tary debate, of a strike, or of a well-regulated negotiation. 

It is important to realize that this conceptual decision is not merely 
of terminological significance. It implies, and is supposed to imply, 
that civil war and parliamentary debate, strike and negotiation are 
essentially motivated by the same type of social relationship and are 
therefore but different manifestations of an identical force. In what 
sense this definition of conflict makes possible a fruitful reformulation 
of problems will become apparent in our subsequent considerations. 
We have already seen, however, that by identifying conflict and revo- 

136 Sociological Critique of Marx 

lution, or conflict and civil war, Marx has obscured more problems 
than he solved. Whoever uses the category of class without assuming 
the presence of class conflict abuses this category. It is the declared aim 
of class theory to explain one type of constitutional group conflict in 
social structures. But the empirical hypothesis is false that insists that 
this class conflict must always assume the form of violent civil war and 
"class struggle." Indeed, it seems plausible that under certain condi- 
tions (which it is possible to determine) class antagonism becomes 
latent or is reactivated from a state of latency. Social classes and class 
conflict are categories connected inseparably j but type and intensity 
of the conflicts in which particular classes are involved in a particular 
situation can be discovered only by studying empirical conditions. 


For Marx, the determinant of social classes was effective private 
property in the means of production. In all essential elements, his 
theory of class is based on this definition of the concept of class. We 
have seen, meanwhile, that precisely this tie between the concept of 
class and the possession of, or exclusion from, effective private prop- 
erty limits the applicability of class theory to a relatively short period 
of European social history. A theory of class based on the division of 
society into owners and nonowners of means of production loses its 
analytical value as soon as legal ownership and factual control are 
separated. For this reason, any effective supersedure of Marx's theory 
of class has to start at this point. Now, it is one of the central theses 
of this study that such a supersedure is possible if we replace the 
possession, or nonpossession, of effective private property by the exer- 
cise of, or exclusion from, authority as the criterion of class forma- 
tion. Renner, Schumpeter, Burnham, Djilas, and others have pre- 
pared the ground for this decision; by contrast to most of these we 
shall not confine the notion of authority to the control of the means 
of production, but consider it as a type of social relations analytically 
independent of economic conditions. The authority structure of entire 
societies as well as particular institutional orders within societies (such 
as industry) is, in terms of the theory here advanced, the structural 
determinant of class formation and class conflict. The specific type 
of change of social structures caused by social classes and their con- 
flicts is ultimately the result of the differential distribution of posi- 
tions of authority in societies and their institutional orders. Control 
over the means of production is but a special case of authority, and the 
connection of control with legal property an incidental phenomenon 

Sociological Critique of Marx 137 

of the industrializing societies of Europe and the United States. 
Classes are tied neither to private property nor to industry or eco- 
nomic structures in general, but as an element of social structure and 
a factor effecting change they are as universal as their determinant, 
namely, authority and its distribution itself. On the basis of a con- 
cept of class defined by relations of authority, a theory can be formu- 
lated which accounts for the facts described by Marx as well as for 
the changed reality of post-capitalist society. 

At several points of our investigation it has become apparent how 
many doubts and objections can be raised against Marx's treatment 
of the relationship between property and social class. In presenting 
Marx's theory, in describing the phenomenon of the separation of 
ownership and control, and in discussing Burnham's inferences from 
this phenomenon and Djilas's analysis of Communist totalitarianism, 
we have seen how, by connecting the concept of class with private 
property (and thereby capitalism), Marx renders this concept fit for 
inclusion in his philosophical conception of history but unfit for the 
sociological analysis even of the conflicts with which he was concerned. 
Marx, too, is concerned with relations of authority j indeed, he ex- 
plicitly refers to these when he describes class conflicts generated by 
the structure of the industrial enterprise. But Marx believed that 
authority and power are factors which can be traced back to a man's 
share in effective private property. In reality, the opposite is the case. 
Power and authority are irreducible factors from which the social 
relations associated with legal private property as well as those asso- 
ciated with communal property can be derived. Burnham, and above 
all Geiger, have rightly stressed that property is in its sociological 
aspect in the first place a permission to exclude others from control 
over an object. It is therefore (Weber, 33*^, p, 28 ) a "chance to find 
obedience with defined persons for an order" (in this case a prohi- 
bition), i.e., a form of authority. But property is by no means the 
only form of authority j it is but one of its numerous types. Whoever 
tries, therefore, to define authority by property defines the general 
by the particular — an obvious logical fallacy. Wherever there is 
property there is authority, but not every form of authority implies 
property. Authority is the more general social relation. 

This formal argument is not, however, the only reason for sub- 
stituting for Marx's definition of classes by private property one by 
a man's share in authority j this generalization is necessary, also, for 
the sake of the empirical applicability of the theory of class. For this 
purpose, it is moreover necessary to separate radically the concept of 

138 Sociological Critiqtie of Marx 

authority from its narrow application to the control over economic 
means of production. Just as property is formally, thus control over 
the means of production is empirically but a special case of those gen- 
eral relations of authority which, according to our conception, lie at the 
base of class formation and class conflict. Why this extension is em- 
pirically necessary will be shown in detail in the following section 
of this chapter, where we deal with the relation between industrial 
and social authority structures. However, this much can be stated 
even without a more detailed discussion: that a theory of group conflict 
the central category of which is defined by a man's share in the control 
of the means of production can apply only to the sphere of industrial 
production. In any case, its significance for structure change would 
be even more restricted than is the significance of the theory of class. 

To say that classes are based on a man's share in legitimate power 
is not to formulate an empirical hypothesis. If this were so, it would 
presuppose an independent definition of the concept of class. It is 
rather a definition which, in a preliminary way, we can state as fol- 
lows: classes are social conflict groups the determinant (or differentia 
specifica) of which can be found in the participation in or exclusion 
from the exercise of authority within any imperatively coordinated 
association. In this sense, classes differ from other conflict groups 
which rest on religious, ethnic, or legal differences. In principle, a 
definition is of course an arbitrary decision. If it is logically unassail- 
able, it cannot be refuted by empirical facts. Yet the definition pro- 
posed here is more than a terminological decision without empirical 
consequences. We shall see that this decision alone opens up many 
new possibilities for the analysis of social conflicts. 

At the same time, the definition of classes by people's participa- 
tion in or exclusion from the exercise of authority distinguishes this 
category clearly from many earlier definitions.^^ The concept of class 
proposed here as promising an eflPective supersedure of Marx's con- 
cept is not based on the level or source of income. Even those sociolo- 
gists who rightly sought the analytical place of a theory of class in 
the study of social conflict have tended to retain two aspects of the 
Marxian concept of class which I propose to abandon. Most of them 

^^ It is perhaps necessary to emphasize at this point that our definition, as formu- 
lated so far, contains many ambiguities and is bound to raise doubts. In this chapter, 
I confine myself to statements required by the critical dissociation from Marx. A more 
detailed discussion of power and authority as determinants of social class will be found 
in the subsequent chapter; in the course of this discussion the definition proposed here 
will become rather more specific and, I hope, unambiguous. 

Sociological Critique of Marx 139 

maintained (i) that classes are in some sense "economic** groupings 
and (2) that the lines of class structure run parallel to those of social 
stratification. So far as I can see, neither Marshall nor Geiger, neither 
Schumpeter nor Lipset and Bendix have renounced these two stipula- 
tions. "To us," Lipset and Bendix state (55, pp. 244, 248), "the pur- 
pose of research in this field is an analysis of the incessant interplay 
between the factors of stratification which make for social change and 
those which tend to arrest it." And: "The analysis of social class is 
concerned with an assessment of the chances that common economic 
conditions . . . will lead to organized action." Geiger similarly views 
social classes as "a special case of social strata/^ namely, strata "de- 
termined by the relations of production" (46, p. 35). Although 
Marshall goes one step further by trying to dissociate the concept of 
class from economic conditions and extend it to a notion of "social 
class," he, too, insists that the phenomenon of class "represents a 
hierarchical social stratification'* (58, p. 90).^^ Both the stipulation 
that class is a phenomenon of stratification and that it is associated with 
economic conditions are fetters from which we have to free this cate- 
gory in order to transform it into a useful tool of social analysis. 

If we define classes by relations of authority, it is ipso facto evi- 
dent that "economic classes," i.e., classes within economic organiza- 
tions, are but a special case of the phenomenon of class. Furthermore, 
even within the sphere of industrial production it is not really eco- 
nomic factors that give rise to class formation, but a certain type of 
social relations which we have tried to comprehend in the notion of 
authority. Classes are neither primarily nor at all economic groupings. 

It is less easy to determine the relation between classes as authority 
groups and the system of social stratification. In the first place, it is 
important to realize that there is no one-to-one correlation between 
class structure and social stratification in the sense that classes result 
from people's place in the hierarchy of stratification. The analyses 
of class and of social stratification are essentially independent subjects 
of sociological inquiry. On the other hand, there is between them a 
significant indirect connection which results from the fact that au- 
thority, the determinant of class, is at the same time one of the 
determinants of social status. It can be demonstrated that there is an 
empirical tendency for the possession of authority to be accompanied, 
within certain limits and with significant exceptions, by high income 
and high prestige, and, conversely, for the exclusion from authority 

^" All italics in this paragraph are mine. 

140 Sociological Critique of Marx 

to be accompanied by relatively low income and prestige. Indeed, it 
is one of the distinguishing features of authority that it can become 
an instrument for the satisfaction of other desires and needs and for 
the attainment of directly gratifying social rewards. Thus, there is 
in most societies a tendential, if not unequivocal, correlation between 
the distribution of authority and the system of social rewards that 
underlies stratification.^^ In this sense, but only in this sense, the 
partial parallelism between the lines of class division and those of 
social stratification may be an empirical fact. One might go further 
and regard this parallelism as probable, as it could be argued that a 
certain correspondence between people's share in authority and in 
social rewards in general is a functional imperative of relatively stable 
societies. But no parallelism between structures of class and stratifi- 
cation can be postulated. Classes can be identical with strata, they can 
unite several strata within them, and their structure can cut right 
through the hierarchy of stratification. 

For purposes of clarity it seemed advisable to state, in the strong- 
est possible terms, the way in which class is independent of property, 
economic conditions, and social stratification. In the abstract, no 
qualification need be made to this statement. Fortunately, however, 
empirical conditions do not usually reproduce the simplicity of our 
assumptions and theories. Although the idea of property, of the re- 
lationships that have to do with production, and of the hierarchy of 
social stratification is, in each instance, clearly distinct from the idea 
of class, these factors have a great deal to do with the realities of 
social class and class conflict. Without doubt, the fact that at the time 
Marx wrote there were capitalists who simultaneously owned and con- 
trolled their enterprises contributed greatly to the formation of classes 
and the antagonism between them. Similarly, the fact that it is pos- 
sible to identify the powerful with the wealthy cannot be overlooked 
in class analysis. While the connection between property and social 
class is not one of definition or mutual dependence, it is one that affects 
the empirical course of class conflict. If distinctions of property are 
superimposed on distinctions of class, class conflict is likely to be more 
violent than if these two lines of social differentiation diverge. An 
analogous argument could be made for class and social stratification. 
In fact, this is one of many points at which Marx has transformed a 

^^ We can leave open, here, whether this correlation can be explained in terms of 
a common basic factor, such as a "value system" (Parsons), or whether it is due to the 
direct eflFect of authority. I doubt that a final answer to this problem is methodolog- 
ically possible. 

Sociological Critique of Marx 141 

correct empirical observation into a false and useless assumption by 
arbitrarily generalizing what was characteristic only of the compara- 
tively short historical period which he lived to see. 


The substitution of relations of authority for those of production 
in defining class is but a radical interpretation of some of the theories 
discussed in the preceding chapter. Djilas, Schumpeter, Renner, Gei- 
ger, and, above all, Burnham in his theory of managerial power, have 
paved the way for this step. But Burnham makes a curious mistake 
which is worth examining in some detail. There is an interesting 
nuance peculiar to his approach which rapidly turns into a consequen- 
tial fallacy and renders his theory empirically nonsensical and ana- 
lytically useless. Burnham tries to supersede Marx^s theory by re- 
placing the narrow legal concept of property by a wider sociological 
concept. Quite rightly he defines property relations (the particular) 
by authority relations (the general). But with a theoretical inaccu- 
racy which is characteristic of his work he now reverses this definition 
and declares authority relations (the general) to be property relations 
(the particular). The managers have property ownership because 
they have factual control. At best, this reversal results in a nonsensi- 
cal extension of the concept of property to all forms of authority, in 
which case the head of state would have property in "his" state. At 
worst, however, and this is Burnham's case, the logical somersault is 
followed by an empirical salto mortale consisting in the assertion that 
authority can exist only where there is property, or, as Burnham says 
himself, that "the instruments of production are the seat of social 
domination" ( 140, p. 125). Marx and Burnham meet in the premise 
that economic power is eo ipso political power, because there is no 
power except that based on ownership in the means of production. 
But both of them are wrong, and their error makes it necessary to pose 
the problem of the relation between economic and social power anew. 

It must be emphasized, in the first place, that the relations be- 
tween industry and society can be established only by empirical in- 
vestigation. There is no axiomatic identity between the managers or 
capitalists of industry and the ministers or highest civil servants of 
the state, just as the exclusion of industrial workers from top political 
positions is by no means an unchangeable element in the structure of 
industrial societies. Again, it was true in Marx's own time, and in 
English society, that the captains of industry or their relatives tended 
to monopolize many of the leading political positions. The same is 

142 Sociological Critique of Marx 

still true In several countries, including the United States and Ger- 
many. But this particular observation does not legitimize the formu- 
lation of a general law. Should this law nevertheless be advanced as 
a hypothesis, then it was refuted by the first government of a labor 
party in an industrial country. The political state and industrial pro- 
duction are two essentially independent associations in which power 
is exercised j and their interrelations are a subject for empirical re- 

Once again, we must supplement, if not correct, an unequivocal 
analytical distinction by empirical facts. A number of empirical gen- 
eralizations about the particular — and especially close — relation be- 
tween the organizations of industrial production and the state are in 
fact possible for modern societies. Among all imperatively coordi- 
nated associations apart from the state, those of industry occupy, in- 
deed, a place which endows them with particular significance in con- 
nection with conflicts in industrial societies. This is above all because 
of three factors which distinguish industry from all other institutional 
orders of society, with the exception of the state itself: the mere size 
of industrial production, its significance for the lives of those who 
participate in it, and the severity of the sanctions at the disposal of 
the rulers of industry. 

If in this context we refer to industry, we always mean primarily 
that sector of the economy of advanced societies which is concerned 
with commodity production in enterprises (factories), and in which 
"means of production" exist in the strict sense of the term. By in- 
dustry we mean, in other words, what C. Clark would call "secondary 
production." This delimitation seems justified by the fact that the 
enterprise of industrial production displays with particular clarity the 
traits of an association coordinated by relations of authority. In in- 
dustrial societies, the sector of secondary production is prominent even 
in terms of its purely physical extent. Nearly one out of every two 
citizens of such societies earns his living in industrial enterprises of 
production. Their position in the national economy exceeds that of all 
other branches of activity. Moreover, the enterprises of production 
themselves grow into mammoth organizations with a hundred thou- 
sand employees or more. Without doubt, Drucker is right in empha- 
sizing the special place of the modern large corporation. 

The particular significance of large-scale industry is further em- 
phasized by the fact that those who earn their living in industrial 
enterprises spend a large part of their lives there and for an even larger 
part are under the influence of the social relations characteristic of 

Sociological Critique of Marx 143 

industry. Sociologists have often emphasized the importance of occu- 
pational roles in industrial societies j these roles also have their rami- 
fications in problems of class conflict. Since the authority relations of 
industrial production occupy so large a space in the lives of so many 
people, they tend to overshadow the authority relations of most, if 
not all, other associations. Except for the political society — the state 
— no other imperatively coordinated organization can compare with 
industrial production in the number of persons affected by its struc- 
ture and in the intensity of this influence. Finally, the special position 
of industrial production is due to the type of sanction inflicted by it. 
Weber defined the state in terms of the monopoly of physical force 
in a given territory. But even in recent times there are examples where 
the managers of industrial enterprises of production have broken this 
monopoly and used their own police force to try to enforce the obedi- 
ence of their workers.^* Even apart from extreme cases of this kind, 
dismissal and even removal to a worse-paid position constitutes in- 
fringements on the lives of people so severe that one could call them 
at least quasi-physical sanctions. The severity of sanctions is not the 
least cause of the fact that under certain conditions conflicts within 
industrial organizations may transcend their limits and dominate the 
scene of social conflict. 

These are empirical generalizations. Although they underline the 
significance of industry for industrial societies, no universal law of 
connection or interrelation between industrial and political power can 
be derived from them. On the contrary, we shall argue at a later point 
that the validity even of these generalizations can be restricted 
by social changes and has been restricted by recent developments. 
Nevertheless, we can conclude from the considerations of this section 
that the organizations of industrial production play a prominent part 
in modern societies. While in principle the problem of the relation 
of particular associations to the political state has to be posed anew 
for each association, that of the relation between industry and society 
has de facto a certain primacy in industrial societies. The state is an 
association coordinated by authority relations, and so is industrial pro- 
duction. The questions whether the structures of the one are also 
those of the other, whether the rulers of industry are also directly 
or indirectly those of the state, and whether the powerless of industry 
are also powerless politically stand in the center of any analysis of 

^* Examples are the cases revealed by the La Follctte Committee on Civil Liber- 
ties in the United States of the 1930's, where some industrialists controlled arsenals 
of arms larger than that of the Chicago City Police. 

144 Sociological Critique of Marx 

industrial societies in terms of class. Although we have to reject as 
untenable or, perhaps, as a refuted empirical generalization Marx's 
assertion that political power follows "necessarily" from industrial 
power, his premise reveals — as is so often the case in his work — a 
correct feeling, an instinct, even, for empirically significant relations. 
Even today it is one of the essential tasks of class analysis to explain 
industrial conflict with the model of a class theory and to examine the 
ramifications of industrial conflict for the political process in terms of 
specific hypotheses. As to the result of such an examination, we can 
but suspect that it will bring out more complex and less one-sided 
relations than Burnham or Marx believed existed. 


The problem of the relation between individuals and classes has 
been as conspicuously neglected by Marx as it has been overempha- 
sized by modern sociologists. It is not surprising that neglect cannot 
solve the problem, but it is unfortunately true that neither has over- 
emphasis. Despite many a treatment of the problem, there is as yet 
no clear formulation of its several dimensions. "For the individual," 
Schumpeter says, "his belonging to a certain class is a given fact"j 
he is "born into a certain class position." More accurately, this does 
not hold for the individual qua individual: "The family, and not the 
physical person, is the true individual of class theory" (27, p. 158). 
Marshall refers to this statement of Schumpeter's but contrasts it with 
the latter's other thesis, according to which "classes which by their 
character and relative position might be called identical social indi- 
viduals never consist of the same family individuals for any length 
of time, but always of different ones" (27, p. 170). Against this, 
Marshall postulates "some permanence in the grouping, so that a 
man who belongs to a certain class remains in it unless — to use a col- 
loquialism — 'something is done about it'" (58, p. 91). Again, we 
encounter the problem in Renner's analysis, which goes even further 
than Marshall's: "Like its causes, class position is almost without ex- 
ception lasting, comprising the whole life and sequence of genera- 
tions. . . . Every class develops in its members a uniform type" 
(70, p. 103). Parsons lays less stress on the time aspect but, like 
Schumpeter, defines the "class position of an individual" as the po- 
sition "which he shares with other members in an effective kinship 
unit," i.e., as a family position (67, p. 328). These quotations are 
selected at random j it would not be difiicult to supplement them with 

Sociological Critique of Marx 145 

others. I quote them only to indicate a general problem about which 
some agreement will have to be reached in order for concept and 
theory of class to become useful tools of sociological analysis. 

So far, we have referred to classes only in connection with social 
positions or roles. Authority structures as well as the associations in 
which they prevail can in principle be analyzed independent of the 
actions and motives of their specific human representatives. They are 
facts of structure which, like the parts of a play or an organization 
chart, can be analyzed without reference to the specific individuals 
who occupy the positions. It seems to me one of the most important 
discoveries of modern sociology that in analyzing class structure — as 
of course most other phenomena of social structure — we can and must 
concentrate on such quasi-objective facts, on roles and role structures. 
Here, again, Marx displayed an admirable instinct. "We are con- 
cerned here," he says in the preface to his Capital, "with persons only 
in so far as they are the personifications of economic categories, car- 
riers of certain class relations and interests" (12, p. 8). However, 
today we should substitute for the "only" in Marx's statement an "in 
the first place," for the second and equally important step of struc- 
tural analysis is concerned with the relation between social roles and 
their personnel. Marx deliberately avoided this stepj he confined 
himself to introducing, wherever necessary, certain ad hoc assumptions 
which bear on this relation. Here, the traditional theory of class re- 
quires not so much criticism or supersedence as supplementation. 

The problem of the relation between social roles and the persons 
who occupy them appears in more recent literature on the theory of 
class under four main aspects which it is useful to distinguish : ( i ) the 
problem of the determinant of class, (2) the problem of class be- 
havior, (3) the problem of stability, and (4) the problem of recruit- 
ment to classes. The first of these problems results, as we shall see, 
from an erroneous conception of the subject matter of class analysis; 
it is in this sense a false problem. The substance of the solution of 
the second and third problems is a matter of class theory itself and 
will therefore be postponed to subsequent chapters j these preliminary 
considerations aim only at the precise formulation of questions re- 
lating to these problems. The solution of the fourth problem follows 
from our earlier discussions and can therefore be formulated here. 

(i) For more than thirty years the distinction between "subjec- 
tive" or "subjectivist" and "objective" or "objectivist" concepts and 
even theories of class — a distinction that is as unclear as it is superflu- 
ous — has deflected sociological discussion from the proper field of 

146 Sociological Critique of Marx 

class analysis. It is little surprising that in this discussion the rubber 
terms "objective" and "subjective" were themselves subject to many 
a fluctuation of meaning. If we return here briefly to this question, 
we do so only because it might appear to have something to do with 
the relation between (subjective?) individuals and (objective?) social 

By "subjective" theories of class most classifiers mean conceptions 
which are in some sense "social psychological" (Geiger), according 
to which "a man belongs to the class that he feels he belongs to" 
(Marshall, 58, p. 93) and which "seek the cause of social classes 
entirely Vithin' the members of a class, in their psyche and values" 
(Croner, 129, p. 154). Analogously, "objective" theories try "to 
determine the basis of class exclusively by ^objective' data, i.e., by 
data which are given in the environment, the conditions of existence, 
etc., of the members of classes" (Croner, 129, p. 148) j such theories 
"represent class as automatically determined by definite criteria, es- 
pecially wealth and occupation" (Marshall, 58, p. 93). Thus, this is 
supposed to be a classification of theories of class according to the 
determinant of social class j it is asserted that this determinant can be 
found either within the individual class members or in conditions 
outside them. At first sight, this distinction may appear plausible, but 
upon closer inspection it turns out to be both meaningless and mis- 

Let us look first at the so-called "subjective" theories. Here, 
classes are based on the psyche of individuals. Croner quotes Centers, 
for whom "class . . . can well be regarded as a psychological phe- 
nomenon in the fullest sense of the term" (38, p. 27). If, therefore, 
a class exists anywhere, this means that people with common or similar 
"psyches" and "values" have found each other. But why do they 
have common "values"? Why do they have a common "class con- 
sciousness"? There are two possible answers to these questions: either 
this individual disposition is in fact an ultimate determinant — in which 
case classes are, from a sociological point of view, random phenomena, 
and there cannot be any theory of class 5 or the class consciousness, 
the "psychology," is in itself a phenomenon generated by and explain- 
able in terms of social structure — in which case there can be a theory 
of class, but it is not "subjective." It is only fair to add that for most 
scholars classified as "subjectivists" psychic phenomena are in fact sec- 
ondary or, which is the same thing, socially structured. Their de- 
terminant of class is in reality not the individual but the social relations 
in which he and others are involved. 

Sociological Critique of Marx 147 

Are, then, all theories of class "objective"? If by "objectivists" 
we mean people who operate with factors like "wealth" or "con- 
ditions of existence," we are dealing, as we know by now, not with 
theorists of class but with theorists of stratification. But if "objective" 
means no more than that class analysis is based on the study of con- 
ditions of social structure, such as relations of authority, then the epi- 
thet is meaningless for two reasons: first, it is hard to see what, if any- 
thing, one says about social structure by calling it "objective" 5 second, 
there never has been, and never can be, a theory of class which does 
not proceed from extra-individual conditions of social structure to 
their individual bearers. 

Indeed, many of the classifiers formulate triumphantly what they 
call "subjective-objective" concepts of class (thus Geiger, Marshall, 
Croner, and others). Theirs is a cheap triumph, for it constitutes no 
more than the solution of a self-made task. They could have saved 
themselves, and us, many words by simply stating that, being soci- 
ologists, they proposed to deal with their subject matter both in its 
structure and in the personnel of this structure. This statement might 
have been a commonplace but it would have done less harm than the 
endless discussion of the spurious problem of "subjectivism" and 
"objectivism. "^° 

(2) The alternative problem whether classes are generated "pri- 
marily" by the structure of social roles or by the psyche of their 
incumbents reveals itself as a senseless construction. If classes are a 
legitimate subject of sociological analysis, their determinant must be 
structural. But there is a genetic problem or, to be more careful, a 
problem of correlation with respect to the relation between social 
classes and the individual personalities of their members. Class con- 
sciousness, community of values, attitudes, "cultures," and behavior 
patterns are no less important for the theory of class just because they 
are assumed to be structured. We have to differentiate, here, between 
the determinant of classes and their empirical character. Empirically, 
of course, classes consist of human individuals. Even if individuals 
are members of classes only as incumbents of certain roles — i.e., with 
a sector of their personalities — problems of attitude and behavior must 
be raised. No theory of class can ignore them. In particular it will 

^^ By way of qualification of this polemical conclusion it is only fair to add that 
there is in fact a difference of emphasis among different theories of class. Thus, it is 
undoubtedly true that Marx has put less emphasis on the psychological aspects of class 
action than Centers. But this differential distribution of emphasis cannot justify any 
inference as to the nature of class theory. 

148 Sociological Critique of Marx 

be necessary to include in the formulation of such a theory certain 
generalizations regarding the questions: {a) which motivational di- 
rectives follow quasi-automatically from the incumbency of social 
roles relevant for class conflict (class interests), {b) under which con- 
ditions these structured directives become conscious motives recog- 
nized by the individual (class consciousness), and (c) by which addi- 
tional common features of primarily psychological reality social classes 
or some of their forms are characterized (class culture). All these 
problems are closely allied with the systematic formulation of class 
theory. We shall therefore have to deal with them in some detail 
in the subsequent chapters. 

(3) The problems of motivation and class behavior are evidently 
closely connected with a further problem which we have encountered 
repeatedly, most recently in the quotations at the beginning of this 
section. I mean the question : how permanent is an individual's mem- 
bership in a class, and how permanent has it to be in order for classes 
to persist and operate as such? Here, sociologists hold widely diver- 
gent opinions. Whereas for Renner "class position is almost without 
exception lasting, comprising the whole life and the sequence of gen- 
erations," for Schumpeter classes "never consist of the same family 
individuals for any length of time." This, too, is a problem in the 
personnel of social classes, the solution of which does not by any means 
follow from the analysis of the structural determinant of class but 
requires, instead, an independent investigation. It touches on the 
problem of social mobility of individuals within and between gen- 
erations. However, as with the psychology of social classes, their 
stability and solidarity can adequately be dealt with only in the con- 
text of the systematic exposition of class theory. We shall return to 
this problem in the subsequent chapters of this study. 

(4) However, the problem of the relation between individuals 
and classes that underlies all others can be solved with the elements 
assembled so far. On the one hand, classes are somehow based on a 
structural arrangement of social roles j on the other hand, they con- 
sist of persons. How do these two get together? How are social classes 
recruited? How does the individual become a member of a class? Are 
people born into classes, or do they acquire membership by achieve- 
ment? These questions, which appear without fail in any analysis of 
social structure, refer to the problem of role allocation or of the re- 
cruitment of the personnel for social roles. They can be answered by 
analogy to general sociological procedure. Class conflict results ulti- 
mately from the distribution of authority in social organizations. 

Sociological Critique of Marx 149 

Classes are based on the differences in legitimate power associated with 
certain positions, i.e., on the structure of social roles with respect to 
their authority expectations. It follows from this that an individual 
becomes a member of a class by playing a social role relevant from 
the point of view of authority.^** Since any role in any social organi- 
zation coordinated by authority is relevant in this sense, and since, 
further, every individual belongs at least to political society, every 
individual belongs by virtue of such membership to at least one class. 
He belongs to a class because he occupies a position in a social organi- 
zation} i.e., class membership is derived from the incumbency of a 
social role. In this sense the criterion of allocation of individuals and 
classes is subordinated to the criterion of the allocation of individuals 
and authority roles. The question "How does the individual become 
a member of the working class?" can be reduced to "How does he 
become a worker?" 

If Schumpeter states in general that the individual "is born into 
a certain class position," this statement is as a general proposition false. 
The individual is born into a class position only in societies in which 
he is born into a position endowed with or deprived of authority. If 
participation in governmental functions, or exclusion from these, is 
hereditary, class membership, too, is hereditary. If, on the other hand, 
admission to positions of authority is based on individual ability or 
achievement, class membership, too, is achieved. The industrial 
worker who, by virtue of a law, is elected an executive of an industrial 
enterprise thereby changes both his authority position and his (in- 
dustrial) class membership. The principle of recruitment to social 
classes must respond to the relevant social structures. The determi- 
nation of class is therefore a task of empirical inquiry and cannot be 
made in general 5 it can be made only for specific societies. 

With some qualifications, this conclusion also holds for the ques- 
tion whether the "physical person" or the family (the "effective kin- 
ship unit") is the "true individual" of a social class. This question 
cannot be decided in general either j it is meaningful only in the con- 
text of specific arrangements of social structure. In a society in which 
the wife, children, and possibly even other relatives of an entrepre- 
neur "borrow" their entire social position from him, can replace him 
in it, and share his social position in this sense at least potentially, the 
(extended) family is indeed the "individual" of a class. But if social 

^^ In the next chapter, this statement will be qualified by the distinction between 
quasi-groups and interest groups. Properly speaking, the present analysis holds for 
quasi-groups only; interest groups presuppose additional factors. 

150 Sociological Critique of Marx 

positions are fundamentally individualized, if wife and children of 
a manager can make no claim to his position when he retires or dies, 
then the "physical person" is also the individual of the class, and it 
is therefore possible for members of the same family to belong not 
only to different classes (classes generated by different imperatively 
coordinated associations) but, indeed, to opposing classes (classes 
opposed within the same association). The manifold empirical types 
of class will become more apparent in the course of our investiga- 
tion j criteria of recruitment give an indication of their possible dif- 


Several early German sociologists have carried on a passionate 
debate about a problem which is not dissimilar to that of "subjectivist" 
and "objectivist" theories of class and has in fact explicitly been 
identified with this problem by Geiger. This problem, too, relates 
to the concept of class and was always formulated as an alternative: 
Are classes a "real phenomenon" {Realfh'dnomen) or a "theoretical 
phenomenon" ( Ordnungsfh'dnomen) ? Are they realities or construc- 
tions of science? "The term class," Geiger says (91, p. 2), "occurs on 
the one hand as the abstract of men of one type^ and on the other hand 
as the concept of a ^o/Z^^j/m/y. . . . In the first case men are classified 
on the basis of certain characteristics or sets of characteristics ['theoreti- 
cal phenomenon* — R.D.] .... The concept of class as a collectivity 
has a different origin. . . . Class in this sense is the concept of a social 
entity which as such involves a specific goal and intention, is the con- 
cept of a specific totality" ['real phenomenon' — R.D.] . Even before 
Geiger, Schumpeter had introduced a similar distinction between class 
as a "particular social creature which acts and suffers as such and wants 
to be understood as such" ['real phenomenon'] and class in the sense 
of "orders of pluralities according to certain characteristics. Under- 
stood in this sense class is a creature of the scientist and owes its ex- 
istence to his ordering hand" ['theoretical phenomenon'] (27, pp. 
149 f.).^^ Both Geiger and Schumpeter emphatically decide in favor 
of an understanding of class as a "real phenomenon" and relegate the 
"theoretical phenomenon" to a lower level of analysis. But both have 
overlooked the fact that they have fallen victim, here, to a false prob- 
lem much like the classifiers of "subjectivism" and "objectivism." 

The concept of class as described so far in this study has, indeed, 

^^ The same "problem" reappears in some recent American discussions of the 
sociolog/ of class, for which cf. Lenski (54). 

Sociological Critique of Marx 1 5 1 

two distinct aspects. On the one hand, we have dealt with classes as 
effective forces in social conflicts, even as organized groupings carry- 
ing on such conflicts. As such, classes are obviously "real phenomena," 
i.e., empirically identifiable "social entities" or "creatures." On the 
other hand, we have derived classes from positions in associations co- 
ordinated by authority and defined them by the "characteristic" of 
participation in or exclusion from the exercise of authority. In this 
sense, classes are evidently "theoretical phenomena," "creatures of the 
scientist," and not organized groupings. There can be no doubt that 
there is a difference between these two "definitions." But is it neces- 
sary to decide in favor of one or the other of these "definitions"? Are 
they really mutually exclusive alternatives.? These questions can be 
answered in the afiirmative only if one is concerned not with a theory 
of class but merely with the formulation of a descriptive category. 
Whoever answers it in the affirmative thereby explicitly renounces the 
development of a theory of class. 

As with the case of alternative "subjective" or "objective" con- 
cepts, the fallacy of the problem derives from the fact that what is 
basically an analytical or genetic problem is projected, so to speak, 
from the third into the second dimension and thereby falsely creates 
what appears to be a logical alternative. One can contrast a caterpillar 
and a butterfly and state triumphantly that they are different, but then 
one must not be surprised to find that a "two-dimensional" treatment 
of this kind does not permit the question as to whether the one may 
have developed out of the other. In the case of classes, no problem of 
genesis in this real sense is involved, but we do find an analogous situa- 
tion on the level of analysis. One may of course confine oneself to 
describing the "real phenomenon" and the "theoretical phenomenon" 
of class as an "alternative of formal-logical possibilities" (Geiger, 91, 
p. 2). But if one does so, it is no longer possible to ask whether the 
structural analysis of the one requires the assumption of the other. 
Analyses, explanations, theories are always "creatures of the scientist," 
and this holds for their elements, too. But can this be regarded as an 
objection? Is it not rather the very point and substance of all science 
to explain "real phenomena" in terms of "theoretical phenomena" by 
dissecting the living richness of the one with the tools of the other and 
reconstructing it on the level of theory? 

Schumpeter and Geiger were indeed on doubtful ground not only 
in constructing an artificial alternative, but also in deciding uncondi- 
tionally for one of its sides. Their choice can at best be accepted and 
considered as a methodological principle, but even as such it is not 

152 Sociological Critique of Marx 

justifiable. It may be useful to start the formulation of theories by 
considering real problems instead of by constructing "reality" out of 
the skies of theory; it may be sensible to derive the general from the 
particular, instead of starting with the general. However, reasonable 
as this procedure may be from the point of view of the psychology of 
scientific discovery, it is erroneous to derive from it a principle of the 
logic of scientific discovery. Logically, at least, a theory takes prece- 
dence over a hypothesis, a hypothesis over a descriptive statement. 
Moreover, it is empirically of no consequence for the validity of a 
theory whether it be formulated with a view to one, ten, or a hundred 
"real phenomena" or, indeed, independent of these in abstracto. What 
matters, rather, is whether and how a theory illuminates its proper area 
of reality, and whether empirical processes refute the hypothesis de- 
rived from the theory. 

Many of the considerations of this chapter were concerned with 
the determinant and context of the concept of class. It was necessary, 
first of all, to clarify the most important prerequisites of a sociological 
theory of class. But as an isolated category the concept of class is mean- 
ingless even for purely descriptive purposes. The statement that the 
managers or bureaucrats of industry constitute an industrial class is 
more than a mere designation, an empty quid -pro quoy only if "class" 
is not merely a defined term but a category embedded in a theory. 
Concept and theory of class are inseparably connected. For this reason, 
the considerations of this chapter have been more than a mere dis- 
cussion or definition of the concept of class; at every stage they pointed 
beyond the category of class into the field of class theory. Before we 
embark on a systematic discussion of this theory, it seems appropriate 
to try and delimit its field a little. 

Class theory is concerned with the systematic explanation of that 
particular form of structure-changing conflict which is carried on by 
aggregates or groups growing out of the authority structure of social 
organizations. The general theory of class precedes the empirical 
analysis of given societies in terms of class in that it states the under- 
lying regularities of class conflict in a form that in principle allows ap- 
plication to all societies. But the following formulation of the theory 
of class does not claim universal applicability, for such applicability is 
always subject to the test of empirical research; it is confined, instead, 
to that type of society which we have described as industrial society. 
Its extension to other types of society may be possible and will in fact 
be suggested at several points; but a thorough discussion of class 
theory on this most general level falls outside the limits of the present 

Sociological Critique of Marx 153 

The general theory of class consists of two analytically separable 
elements: the theory of class formation and the theory of class action, 
or class conflict. Schumpeter does not sufficiently recognize this dis- 
tinction if he differentiates between the problems of the "essence of 
the phenomenon," "class association," "class formation," and the "con- 
crete causes and conditions of an individually specific, historically 
given class structure" (27, p. 151). By the "problem of essence" he 
seems to understand, above all, the problem of definition, which is pre- 
liminary from the point of view of class theory. The problems of 
"class association" ("How and why do classes hang together?") and 
"class formation" ("How do classes originate?") belong together and 
will here be dealt with under the one heading of class formation. The 
problem of concrete empirical conditions of given class structures is 
part of the theory of class only by privation: the theory must make 
clear where its general propositions must be supplemented with em- 
pirical observations and where only empirical generalizations, not 
"laws" or postulates, are possible. Schumpeter does not even formu- 
late the important problem of the regularities of class conflict and the 
relations between classes.^* 

The theory of class formation — which will be dealt with in Chap- 
ter V — is concerned with the question of analyzing the "genesis" of 
social classes. The theory must establish relations which connect the 
specific "real phenomenon" class by way of the "theoretical phenome- 
non" class with patterns of social structure, and in this sense derive 
social classes from social structure. This is evidently a problem of 
genesis, but it will prove useful to use this word in quotes. The ana- 
lytical reduction ("explanation") of social classes to structural condi- 
tions cannot be understood as an empirical generalization of what 
actually happens in the emergence and formation of classes. In mak- 
ing a structural analysis of the class phenomenon we do not assert 
that a given arrangement of structure "necessarily" results in the full 
formation of organized classes, or that every step of analysis reflects 
a factual stage of development in the history of given classes. In so 
far as the theory of class formation is a scientific theory, it can neither 
presuppose, nor imply, nor give rise to empirical generalizations that 
are usually of doubtful logical status. 

The theory of class action^ or class conflict — which will be dealt 
with in Chapter VI — is based on the theory of class formation. Its sub- 
ject matter consists in the general analytical elements of the inter- 

^^ Evidently the reason for this omission is that Schumpeter's concept of class is in 
fact — as we found above — a concept of social stratum. 

154 Sociological Critique of Marx 

relations between classes conceived as structural phenomena. It is con- 
cerned in particular with patterns of class conflict and the regulation 
of class conflict. This aspect of the theory of class appears to come close 
to the limits of the possibility of theoretical analysis. It will indeed be 
our task, here, to determine, in the light of class theory, the area and 
types of variability of empirical classes, class conflicts, and changes 
caused by class conflict, and to define the points at which the theory of 
class has to be supplemented by empirical generalizations. 


Toward a Sociological Theory of Conflict 
in Industrial Society 


Social Structure, Group Interests, and 
Conflict Groups 


Throughout the history of Western political thought, two views 
of society have stood in conflict. Both these views are intended to 
explain what has been, and will probably continue to be, the most 
puzzling problem of social philosophy: how is it that human societies 
cohere? There is one large and distinguished school of thought ac- 
cording to which social order results from a general agreement of 
values, a consensus omnium- or volonte generale which outweighs all 
possible or actual differences of opinion and interest. There is an- 
other equally distinguished school of thought which holds that co- 
herence and order in society are founded on force and constraint, on 
the domination of some and the subjection of others. To be sure, 
these views are not at all points mutually exclusive. The Utopian 
(as we shall call those who insist on coherence by consensus) does not 
deny the existence of differences of interest j nor does the Rationalist 
(who believes in coherence by constraint and domination) ignore such 
agreements of value as are required for the very establishment of 
force. But Utopian and Rationalist alike advance claims of primacy 
for their respective standpoints. For the Utopian, differences of in- 
terest are subordinated to agreements of value, and for the Rationalist 
these agreements are but a thin, and as such ineffective, coating of 
the primary reality of differences that have to be precariously recon- 
ciled by constraint. Both Utopians and Rationalists have shown much 
ingenuity and imagination in arguing for their respective points of 
view. This has not, however, led them more closely together. There 
is a genuine conflict of approach between Aristotle and Plato, Hobbes 
and Rousseau, Kant and Hegel, and this conflict has grown in in- 
tensity as the history of thought has advanced. Unless one believes 
that all philosophical disputes are spurious and ultimately irrelevant, 

158 Social Structure y Grouf Interests y Conflict Groups 

the long history of the particular dispute about the problem of social 
order has exposed — if not solved — what appear to be fundamental 
alternatives of knowledge, moral decision, and political orientation. 

Conflicting philosophical positions must inevitably, it seems to me, 
reappear constantly in theories of science. Even if this should not 
generally be the case, I would claim that the philosophical alternative 
of a Utopian or a Rational solution of the problem of order pervades 
modern sociological thinking even in its remotest manifestations. 
Here, as elsewhere, philosophical positions do not enter into scien- 
tific theories unchanged. Here, as elsewhere, they pass through the 
filter of logical supposition before they become relevant for testable 
explanations of problems of experience. The sociological Utopian 
does not claim that order is based on a general consensus of values, but 
that it can be conceived of in terms of such consensus, and that, if it 
is conceived of in these terms, certain propositions follow which are 
subject to the test of specific observations. Analogously, for the so- 
ciological Rationalist the assumption of the coercive nature of social 
order is a heuristic principle rather than a judgment of fact. But this 
obvious reservation does not prevent the Utopians and the Rational- 
ists of sociology from engaging in disputes which are hardly less 
intense (if often rather less imaginative and ingenious) than those 
of their philosophical antecedents. The subject matter of our concern 
in this study demands that we take a stand with respect to this dispute. 

Twice in our earlier considerations we have been faced with dif- 
ferences in the image of society — as I then called it — which cor- 
respond very closely to the conflicting views of Utopians and Ra- 
tionalists. I have tried to show that, at least in so far as historical 
societies are concerned, Marx subscribed to an image of society of 
the Rational variety. He assumed the ubiquity of change and conflict 
as well as domination and subjection, and I suggest that this view 
seems particularly appropriate for the analysis of problems of con- 
flict. In any case, it seems more appropriate than the Utopian view 
implicit in the works of Drucker and Mayo, according to which happy 
cooperation is the normal state of social life. Marx, or Drucker and 
Mayo, may not be especially convincing representatives of these 
views,^ but the distinction with which we are concerned here is, in 

^ This would be true, of course, for rather difTerent reasons. Drucker and Mayo 
are rather lacking in subtlety, and it is therefore too easy to polemicize against their 
positions. Marx, on the other hand, is certainly subtle, but his notions of the "original" 
and the "terminal" societies of (imaginary) history demonstrate that he was but a 
limited Rationalist with strong Utopian leanings. Such mixtures of views really quite 
incompatible are in fact not rare in the history of social thought. 

Social Structure y Group Interests ^ Conflict Groups 159 

any case, not tied to their names. Generally speaking, it seems to me 
that two (meta-) theories can and must be distinguished in contem- 
porary sociology. One of these, the integration theory of society , con- 
ceives of social structure in terms of a functionally integrated system 
held in equilibrium by certain patterned and recurrent processes. The 
other one, the coercion theory of society ^ views social structure as a 
form of organization held together by force and constraint and reach- 
ing continuously beyond itself in the sense of producing within itself 
the forces that maintain it in an unending process of change. Like 
their philosophical counterparts, these theories are mutually exclu- 
sive. But — if I may be permitted a paradoxical formulation that will 
be explained presently — in sociology (as opposed to philosophy) a 
decision which accepts one of these theories and rejects the other is 
neither necessary nor desirable. There are sociological problems for 
the explanation of which the integration theory of society provides 
adequate assumptions j there are other problems which can be ex- 
plained only in terms of the coercion theory of society j there are, 
finally, problems for which both theories appear adequate. For socio- 
logical analysis, society is Janus-headed, and its two faces are equiva- 
lent aspects of the same reality. 

In recent years, the integration theory of society has clearly domi- 
nated sociological thinking. In my opinion, this prevalence of one 
partial view has had many unfortunate consequences. However, it 
has also had at least one agreeable consequence, in that the very one- 
sidedness of this theory gave rise to critical objections which enable 
us today to put this theory in its proper place. Such objections have 
been stimulated with increasing frequency by the works of the most 
eminent sociological theorist of integration, Talcott Parsons. It is not 
necessary here to attempt a comprehensive exposition of Parsons' 
position J nor do we have to survey the sizable literature concerned 
with a critical appraisal of this position. To be sure, much of this 
criticism is inferior in subtlety and insight to Parsons' work, so that 
it is hardly surprising that the sociological climate of opinion has re- 
mained almost unaffected by Parsons' critics. There is one objection 
to Parsons' position, however, which we have to examine if we are 
to make a systematic presentation of a theory of group conflict. In a 
remarkable essay, D. Lockwood claims "that Parsons' array of con- 
cepts is heavily weighted by assumptions and categories which relate 
to the role of normative elements in social action, and especially to 
the processes whereby motives are structured normatively to ensure 
social stability. On the other hand, what may be called the substratum 
of social action, especially as it conditions interests which are produc- 

1 60 Social Structure y Group Interests , Conflict Groups 

tive of social conflict and instability, tends to be ignored as a general 
determinant of the dynamics of social systems" (210, p. 136). Lock- 
wood's claim touches on the core of our problem of the two faces of 
society — although his formulation does not, perhaps, succeed in ex- 
posing the problem with suflicient clarity. 

It is certainly true that the work of Parsons displays a conspicuous 
bias in favor of analysis in terms of values and norms. It is equally 
true that many of those who have been concerned with problems of 
conflict rather than of stability have tended to emphasize not the nor- 
mative but the institutional aspects of social structure. The work of 
Marx is a case in point. Probably, this difference in emphasis is no 
accident. It is nevertheless as such irrelevant to an understanding of 
or adoption of the alternative images of society which pervade po- 
litical thought and sociological theory. The alternative between "nor- 
mative elements in social action" and a factual "substratum of social 
action," which Lockwood takes over from the work of Renner, in 
fact indicates two levels of the analysis of social structure which are 
in no way contradictory. There is no theoretical reason why Talcott 
Parsons should not have supplemented (as indeed he occasionally 
does) his analysis of normative integration by an analysis of the in- 
tegration of social systems in terms of their institutional substratum. 
However we look at social structure, it always presents itself as com- 
posed of a moral and a factual, a normative and an institutional, level 
or, in the doubtful terms of Marx, a superstructure and a substratum. 
The investigator is free to choose which of these levels he wants to 
emphasize more strongly — although he may be well-advised, in the 
interest of clarity as well as of comprehensiveness of his analysis, not 
to stress one of these levels to the exclusion of the other. 

At the same time, there is an important element of genuine cri- 
tique in Lockwood's objection to Parsons, When Lockwood contrasts 
stability and instability, integration and conflict, equilibrium and dis- 
equilibrium, values and interests, he puts his flnger on a real alter- 
native of thought, and one of which Parsons has apparently not been 
sufficiently aware. For of two equivalent models of society. Parsons 
has throughout his work recognized only one, the Utopian or inte- 
gration theory of society. His "array of concepts" is therefore in- 
capable of coping with those problems with which Lockwood is con- 
cerned in his critical essay, and which constitute the subject matter 
of the present study. 

For purposes of exposition it seems useful to reduce each of the 
two faces of society to a small number of basic tenets, even if this 

Social Structure, Group Interests, Conflict Groups 1 6 1 

involves some degree of oversimplification as well as overstatement. 
The integration theory of society, as displayed by the work of Par- 
sons and other structural-functionalists, is founded on a number of 
assumptions of the following type: 

(1) Every society is a relatively persistent, stable structure of ele- 

(2) Every society is a well-integrated structure of elements. 

(3) Every element in a society has a function, i.e., renders a contri- 
bution to its maintenance as a system. 

(4) Every functioning social structure is based on a consensus of 
values among its members. 

In varying forms, these elements of (i) stability, (2) integration, 
(3) functional coordination, and (4) consensus recur in all structural- 
functional approaches to the study of social structure. They are, to 
be sure, usually accompanied by protestations to the effect that sta- 
bility, integration, functional coordination, and consensus are only 
"relatively" generalized. Moreover, these assumptions are not meta- 
physical propositions about the essence of society 5 they are merely 
assumptions for purposes of scientific analysis. As such, however, they 
constitute a coherent view of the social process' which enables us to 
comprehend many problems of social reality. 

However, it is abundantly clear that the integration approach to 
social analysis does not enable us to comprehend all problems of social 
reality. Let us look at two undeniably sociological problems of the 
contemporary world which demand explanation. ( i ) In recent years, 
an increasing number of industrial and commercial enterprises have 
introduced the position of personnel manager to cope with matters 
of hiring and firing, advice to employees, etc. Why? And: what are 
the consequences of the introduction of this new position? (2) On 
the 1 7th of June, 1953, the building workers of East Berlin put down 
their tools and went on a strike that soon led to a generalized revolt 
against the Communist regime of East Germany. Why? And: what 
are the consequences of this uprising? From the point of view of the 
integration model of society, the first of these problems is susceptible 

' It is important to emphasize that "stability" as a tenet of the integration theory 
of society does not mean that societies are "static." It means, rather, that such processes 
as do occur (and the structural-functional approach is essentially concerned with 
processes) serve to maintain the patterns of the system as a whole. Whatever criticism 
I have of this approach, I do not want to be misunderstood as attributing to it a 
"static bias" (which has often been held against this approach without full considera- 
tion of its merits). 

1 62 Social Structure y Group Interests y Conflict Groups 

of a satisfactory solution. A special position to cope with personnel 
questions is functionally required by large enterprises in an age of 
rationalization and "social ethic" j the introduction of this position 
adapts the enterprise to the values of the surrounding society j its 
consequence is therefore of an integrative and stabilizing nature. But 
what about the second problem? Evidently, the uprising of the 17th 
of June is neither due to nor productive of integration in East Ger- 
man society. It documents and produces not stability, but instability. 
It contributes to the disruption, not the maintenance, of the existing 
system. It testifies to dissensus rather than consensus. The integra- 
tion model tells us little more than that there are certain "strains" 
in the "system." In fact, in order to cope with problems of this kind 
we have to replace the integration theory of society by a different 
and, in many ways, contradictory model. 

What I have called the coercion theory of society can also be 
reduced to a small number of basic tenets, although here again these 
assumptions oversimplify and overstate the case: 

(1) Every society is at every point subject to processes of change j 
social change is ubiquitous. 

(2) Every society displays at every point dissensus and conflict j 
social conflict is ubiquitous. 

(3) Every element in a society renders a contribution to its disinte- 
gration and change. 

(4) Every society is based on the coercion of some of its members 
by others. 

If we return to the problem of the German workers' strike, it will 
become clear that this latter model enables us to deal rather more 
satisfactorily with its causes and consequences. The revolt of the 
building workers and their fellows in other industries can be explained 
in terms of coercion.^ The revolting groups are engaged in a conflict 
which "functions" as an agent of change by disintegration. A ubiqui- 
tous phenomenon is expressed, in this case, in an exceptionally intense 
and violent way, and further explanation will have to account for 
this violence on the basis of the acceptance of conflict and change as 
universal features of social life. I need hardly add that, like the in- 
tegration model, the coercion theory of society constitutes but a set 
of assumptions for purposes of scientific analysis and implies no claim 

^ For purposes of clarity, 1 have deliberately chosen an example from a totali- 
tarian state. But coercion is meant here in a very general sense, and the coercion model 
is applicable to all societies, independent of their specific political structure. 

Social Structure^ Grouf Interests^ Conßict Groups 163 

for philosophical validity — although, like its counterpart, this model 
also provides a coherent image of social organization. 

Now, I would claim that, in a sociological context, neither of these 
models can be conceived as exclusively valid or applicable. They con- 
stitute complementary, rather than alternative, aspects of the struc- 
ture of total societies as well as of every element of this structure. 
We have to choose between them only for the explanation of specific 
problems J but in the conceptual arsenal of sociological analysis they 
exist side by side. Whatever criticism one may have of the advocates 
of one or the other of these models can therefore be directed only 
against claims for the exclusive validity of either.* Strictly speakdng, 
both models are "valid" or, rather, useful and necessary for socio- 
logical analysis. We cannot conceive of society unless we realize the 
dialectics of stability and change, integration and conflict, function and 
motive force, consensus and coercion. In the context of this study, 
I regard this point as demonstrated by the analysis of the exemplary 
problems sketched above. 

It is perhaps worth emphasizing that the thesis of the two faces 
of social structure does not require a complete, or even partial, re- 
vision of the conceptual apparatus that by now has become more or 
less generally accepted by sociologists in all countries. Categories like 
role, institution, norm, structure, even function are as useful in terms 
of the coercion model as they are for the analysis of social integration. 
In fact, the dichotomy of aspects can be carried through all levels of 
sociological analysis j that is, it can be shown that, like social structure 
itself, the notions of role and institution, integration and function, 
norm and substratum have two faces which may be expressed by two 
terms, but which may also in many cases be indicated by an extension 
of concepts already in use. "Interest and value," Radcliffe-Brown 
once remarked, "are correlative terms, which refer to the two sides 
of an asymmetrical relation" (221, p. 199). The notions of interest 
and value indeed seem to describe very well the two faces of the 
normative superstructure of society: what appears as a consensus of 
values on the basis of the integration theory can be regarded as a 
conflict of interests in terms of the coercion theory. Similarly, what 

* This, it seems to me, is the only — if fundamental — legitimate criticism that can 
be raised against Parsons' work on this general level. In The Social System, Parsons 
repeatedly advances, for the integration theory of society, a claim that it is the nucleus 
of "the general" sociological theory — a claim which I regard as utterly unjustified. It 
is Lockwood's main concern also, in the essay quoted above, to reject this claim to uni- 
versal validity. 

1 64 Social Structure) Grouf Interests ^ Conflict Growps 

appears on the level of the factual substratum as integration from the 
point of view of the former model presents itself as coercion or con- 
straint from the point of view of the latter. We shall presently have 
occasion to explore these two faces of societies and their elements 
rather more thoroughly with reference to the two categories of power 
and of role. 

While logically feasible,^ the solution of the dilemma of political 
thought which we have offered here for the more restricted field of 
sociological analysis nevertheless raises a number of serious problems. 
It is evidently virtually impossible to think of society in terms of 
either model without positing its opposite number at the same time. 
There can be no conflict, unless this conflict occurs within a context 
of meaning, i.e., some kind of coherent "system." No conflict is con- 
ceivable between French housewives and Chilean chess players, be- 
cause these groups are not united by, or perhaps "integrated into," a 
common frame of reference. Analogously, the notion of integration 
makes little sense unless it presupposes the existence of different ele- 
ments that are integrated. Even Rousseau derived his volonte gene- 
rale from a modified helium omnium contra omnes. Using one or the 
other model is therefore a matter of emphasis rather than of funda- 
mental difference J and there are, as we shall see, many points at which 
a theory of group conflict has to have recourse to the integration theory 
of social structure. 

Inevitably, the question will be raised, also, whether a unified 
theory of society that includes the tenets of both the integration and 
the coercion models of society is not at least conceivable — for as to its 
desirability there can be little doubt. Is there, or can there be, a 
general point of view that synthesizes the unsolved dialectics of in- 
tegration and coercion? So far as I can see, there is no such general 
model; as to its possibility, I have to reserve judgment. It seems at 
least conceivable that unification of theory is not feasible at a point 
which has puzzled thinkers ever since the beginning of Western 

For the explanation of the formation of conflict groups out of 
conditions of social structure, we shall employ a model that empha- 
sizes the ugly face of society. In the following sections of this chap- 

° As is demonstrated most clearly by the fact that a similar situation can be en- 
countered in physics with respect to the theory of light. Here, too, there are two seem- 
ingly incompatible theories which nevertheless exist side by side, and each of which 
has its proper realm of empirical phenomena: the wave theory and the quantum theory 
of light. 

Social Structure y Group Interests y Conflict Groups 165 

ter I shall try to show how, on the assumption of the coercive nature 
of social structure, relations of authority become productive of clashes 
of role interest which under certain conditions lead to the formation 
of organized antagonistic groups within limited social organizations 
as well as within total societies. By proceeding step by step along these 
lines, we shall eventually be in a position to contrast the rudiments 
of a sociological theory of group conflict with such earlier approaches 
as have been discussed in the first part of this study, and to decide 
whether the category of class is still a useful tool of sociological 


From the point of view of the integration theory of social struc- 
ture, units of social analysis ("social systems") are essentially volun- 
tary associations of people who share certain values and set up insti- 
tutions in order to ensure the smooth functioning of cooperation. 
From the point of view of coercion theory, however, the units of 
social analysis present an altogether different picture. Here, it is not 
voluntary cooperation or general consensus but enforced constraint 
that makes social organizations cohere. In institutional terms, this 
means that in every social organization some positions are entrusted 
with a right to exercise control over other positions in order to ensure 
effective coercion; it means, in other words, that there is a differential 
distribution of power and authority. One of the central theses of 
this study consists in the assumption that this differential distribution 
of authority invariably becomes the determining factor of systematic 
social conflicts of a type that is germane to class conflicts in the tra- 
ditional (Marxian) sense of this term. The structural origin of such 
group conflicts must be sought in the arrangement of social roles 
endowed with expectations of domination or subjection. Wherever 
there are such roles, group conflicts of the type in question are to be 
expected. Differentiation of groups engaged in such conflicts follows 
the lines of differentiation of roles that are relevant from the point 
of view of the exercise of authority. Identification of variously 
equipped authority roles is the first task of conflict analysis j" concep- 

^ To facilitate communication, I shall employ in this study a number of abbrevia- 
tions. These must not however be misunderstood. Thus, "conflict analysis'' in this 
context stands for "analysis of group conflicts of the class type, class being understood 
in the traditional sense." At no point do I want to imply a claim for a generalized 
theory of social conflict. 

1 66 Social Structure y Group Interests y Conflict Groups 

tually and empirically all further steps of analysis follow from the 
investigation of distributions of power and authority. 

"Unfortunately, the concept of power is not a settled one in the 
social sciences, either in political science or in sociology" (Parsons, 
201, p. 139). Max Weber (33), Pareto (25), Mosca (24), later 
Russell (203), Bendix (126), Lasswell (200), and others have ex- 
plored some of the dimensions of this category; they have not, how- 
ever, reached such a degree of consensus as would enable us to employ 
the categories of power and authority without at least brief concep- 
tual preliminaries. So far as the terms "power" and "authority" and 
their distinction are concerned, I shall follow in this study the useful 
and well-considered definitions of Max Weber. For Weber, power 
is the "probability that one actor within a social relationship will be 
in a position to carry out his own will despite resistance, regardless of 
the basis on which this probability rests" 3 whereas authority (Herr- 
schaft) is the "probability that a command with a given specific con- 
tent will be obeyed by a given group of persons" (33^, p. 28). The 
important difference between power and authority consists in the fact 
that whereas power is essentially tied to the personality of individuals, 
authority is always associated with social positions or roles. The 
demagogue has power over the masses to whom he speaks or whose 
actions he controls j but the control of the officer over his men, the 
manager over his workers, the civil servant over his clientele is au- 
thority, because it exists as an expectation independent of the specific 
person occupying the position of officer, manager, civil servant. It is 
only another way of putting this difference if we say — as does Max 
Weber — that while power is merely a factual relation, authority is 
a legitimate relation of domination and subjection. In this sense, 
authority can be described as legitimate power. 

In the present study we are concerned exclusively with relations 
of authority, for these alone are part of social structure and therefore 
permit the systematic derivation of group conflicts from the organi- 
zation of total societies and associations within them. The significance 
of such group conflicts rests with the fact that they are not the product 
of structurally fortuitous relations of power but come forth wherever 
authority is exercised — and that means in all societies under all his- 
torical conditions. ( i ) Authority relations are always relations of 
super- and subordination. (2) Where there are authority relations, 
the superordinate element is socially expected to control, by orders 
and commands, warnings and prohibitions, the behavior of the sub- 
ordinate element. (3) Such expectations attach to relatively perma- 

Social Structure y Group Interests j Conflict Growps 167 

nent social positions rather than to the character of individuals j they 
are in this sense legitimate. (4) By virtue of this fact, they always 
involve specification of the persons subject to control and of the 
spheres within which control is permissible/ Authority, as distinct 
from power, is never a relation of generalized control over others. 
(5) Authority being a legitimate relation, noncompliance with au- 
thoritative commands can be sanctioned j it is indeed one of the func- 
tions of the legal system (and of course of quasi-legal customs and 
norms) to support the effective exercise of legitimate authority. 

Alongside the term "authority," we shall employ (and have em- 
ployed) in this study the terms "domination" and "subj ection." These 
will be used synonymously with the rather clumsy expressions "en- 
dowed with authority" or "participating in the exercise of authority" 
(domination), and "deprived of authority" or "excluded from the 
exercise of authority" (subjection). 

It seems desirable for purposes of conflict analysis to specify the 
relevant unit of social organization in analogy to the concept of social 
system in the analysis of integration. To speak of specification here 
is perhaps misleading. "Social system" is a very general concept 
applicable to all types of organization j and we shall want to employ 
an equally general concept which differs from that of social system by 
emphasizing a different aspect of the same organizations. It seems to 
me that Max Weber's category "imperatively coordinated association" 
{Herrschaftsverband) serves this purpose despite its clumsiness.* 

In conflict analysis we are concerned inter alia with the genera- 
tion of conflict groups by the authority relations obtaining in impera- 
tively coordinated associations. Since imperative coordination, or au- 
thority, is a type of social relation present in every conceivable social 

"^ This element of the definition of authority is crucial. It implies that the man- 
ager who tries to control people outside his firm, or the private lives of people inside 
his firm, trespasses the borderline between authority and power. Although he has 
authority over people in his firm, his control assumes the form of power as soon as it 
goes beyond the specified persons and spheres of legitimate control. This type of 
trespassing is of course frequent in every authority relation; and an empirical phenome- 
non well worth investigating is to what extent the fusion of authority and power tends 
to intensify group conflicts. 

* Parsons, in his translation of Weber's Wirtschajt und Gesdlschajt, suggests "im- 
peratively coordinated group." Any translation of Weber's term is bound to be some- 
what awkward, but it seems to me that the word "group" in Parsons' translation is 
false. Weber uses Verband, e.g., to describe the state, or a church — units of organiza- 
tion which can hardly be called "groups." "Association" is probably as precise an Eng- 
lish equivalent of Verband as is likely to be found. 

1 68 Social Structure, Group Interests, Conflict Groups 

organization, it will be sufficient to describe such organizations simply 
as associations. Despite prolonged terminological discussions, no gen- 
eral agreement has been attained by sociologists on the precise mean- 
ing of the categories "organization," "association," and "institution." 
If I am not mistaken in my interpretation of the trend of termino- 
logical disputes, it appears justifiable to use the term "association" 
in such a way as to imply the coordination of organized aggregates 
of roles by domination and subjection. The state, a church, an enter- 
prise, but also a political party, a trade union, and a chess club are 
associations in this sense. In all of them, authority relations exist; 
for all of them, conflict analysis is therefore applicable. If at a 
later stage we shall suggest restriction to the two great associa- 
tions of the state and the industrial enterprise, this suggestion is 
dictated merely by considerations of empirical significance, not logi- 
cal (or definitional) difference. In looking at social organizations 
not in terms of their integration and coherence but from the point 
of view of their structure of coercion and constraint, we regard them 
as (imperatively coordinated) associations rather than as social sys- 
tems. Because social organizations are also associations, they generate 
conflicts of interest and become the birthplace of conflict groups. 

I have assumed in the preceding remarks that authority is a char- 
acteristic of social organizations as general as society itself. Despite the 
assertion of Renner — and other modern sociologists — that in some 
contemporary societies the exercise of authority has been eliminated 
and replaced by the more anonymous "rule of the law" or other non- 
authoritative relations, I should indeed maintain that authority is a 
universal element of social structure. It is in this sense more general 
than, for example, property, or even status. With respect to post- 
capitalist industrial society, I hope to establish this position more un- 
ambiguously in the final chapters of this study. Generally speaking, 
however, the universality of authority relations would seem evident 
as soon as we describe these relations in a "passive" rather than in an 
"active" sense. Authority relations exist wherever there are people 
whose actions are subject to legitimate and sanctioned prescriptions 
that originate outside them but within social structure. This formula- 
tion, by leaving open who exercises what kind of authority, leaves little 
doubt as to the omnipresence of some kind of authority somehow ex- 
ercised. For it is evident that there are many forms and types of 
authority in historical societies. There are differences of a considerable 
order of magnitude between the relations of the citizen of classical 
Athens and his slaves, the feudal landlord and his villeins and serfs, 
the nineteenth-century capitalist and his workers, the secretary of a 

Social Structure y Group Interests, Conflict Groups 1 69 

totalitarian state party and its members, the appointed manager of a 
modern enterprise and its employees, or the elected prime minister of 
a democratic country and the electorate. No attempt will be made 
in this study to develop a typology of authority. But it is assumed 
throughout that the existence of domination and subjection is a com- 
mon feature of all possible types of authority and, indeed, of all pos- 
sible types of association and organization. 

The notion of power and authority employed in the present study 
represents what Parsons in a critical review of C. W. Mills's book on 
the American power elite {62) calls the "zero-sum" concept of author- 
ity. Parsons objects to this concept, and his argument provides a wel- 
come opportunity to clarify our notion somewhat further and relate 
it to the two models distinguished above. "The essential point at 
present is that, to Mills [and of course to us in this study — R.D.], 
power is not a facility for the performance of function in and on behalf 
of the society as a system, but is interpreted exclusively as a facility for 
getting what one group, the holders of power, wants by preventing 
another group, the *outs,' from getting what it wants" (201, p. 139). 
This statement is unobjectionable, and in so far as Mills really uses 
power "exclusively" in the "zero-sum" sense, I should tend to agree 
also with Parsons' critique. But then Parsons continues, in the same 
passage, to make the same mistake in the opposite direction, and to 
make it deliberately and consideredly: "What this conception does is 
to elevate a secondary and derived aspect of a total phenomenon into 
the central place" [italics mine] . Not surprisingly, Parsons continues 
to point out what is presumably the primary and original aspect of the 
total phenomenon: "It is the capacity to mobilize the resources of the 
society for the attainment of goals for which a general 'public' commit- 
ment has been made, or may be made. It is mobilization, above all, of 
the action of persons and groups, which is binding on them by virtue of 
their position in the society" ( 20 1 , p. 1 40) . A clearer exposition of the 
two faces of society, and of the untenable and dangerous one-sidechiess 
of Parsons' position, is hardly conceivable. 

It is certainly true that for many purposes of analysis, power or — 
as I should prefer to say — authority, both realizes and symbolizes the 
functional integration of social systems. To use a pertinent illustra- 
tion: in many contexts, the elected president or prime minister of 
democratic countries'* represents his country as a whole j his position 

^ This illustration is unambiguous with respect to the president of the United 
States. Elsewhere, the representative and the governmental functions are usually 
separated; in these cases I mean not the head of state (king, president), but the chief of 
government (prime minister, chancellor). 

1 70 Social Structure y Group Interests y Conflict Groups 

expresses therefore the unity and integration of a nation. In other 
contexts, however, the chief of government is but the representative 
of the majority party, and therefore exponent of sectional interests. 
I suggest that as in the position of the prime minister neither of these 
elements is primary or secondary, thus neither the integrative nor the 
disruptive aspect of authority in social analysis is primary or secondary. 
Like all other elements of social structure, authority has two faces — 
those, so to speak, of Mills and of Parsons — and on the highest level 
of abstraction it is illegitimate to emphasize either of these to the ex- 
clusion of the other. Authority is certainly not only productive of con- 
flict j but neither is it only (or even primarily) "a facility for the per- 
formance of function in and on behalf of the society as a system." If 
we are concentrating in this study on what Parsons would call the 
"negative functions" of authority, we do so because this aspect is more 
appropriate and useful for the analysis of structurally generated syste- 
matic social conflicts. 

In referring to the ugly face of authority as a "zero-sum" concept, 
Parsons brings out one further aspect of this category which is es- 
sential for our considerations. By zero-sum, Parsons evidently means 
that from the point of view of the disruptive "functions" of authority 
there are two groups or aggregates of persons, of which one possesses 
authority to the extent to which the other one is deprived of it.^° This 
implies — for us, if not for Parsons — that in terms of the coercion 
theory of society we can always observe a dichotomy of positions in 
imperatively coordinated associations with respect to the distribution 
of authority. Parsons, in his critique of Mills, compares the distribu- 
tion of authority to the distribution of wealth. It seems to me that 
this comparison is misleading. However unequally wealth may be 
distributed, there always is a continuum of possession ranging from 
the lowest to the highest rank. Wealth is not and cannot be conceived 
as a zero-sum concept. With respect to authority, however, a clear line 
can at least in theory be drawn between those who participate in its 
exercise in given associations and those who are subject to the authori- 
tative commands of others. Our analysis of modern societies in later 
chapters will show that empirically it is not always easy to identify the 

^'^ There is one implication of the expression "zero-sum" which would be con- 
trary to my thesis. Mathematically, it would be possible for both groups to have no 
authority in the sense of a complete absence of authority. I have argued above that 
under all conditions the authority of one aggregate is, so to speak, greater than zero, 
and that of the other aggregate correspondingly smaller than zero. The presence of 
authority, and its unequal distribution, are universal features of social structure. 

Social Structure, Group Interests, Conflict Groups 171 

border line between domination and subjection. Authority has not 
remained unaffected by the modern process of division of labor. But 
even here, groups or aggregates can be identified which do not partici- 
pate in the exercise of authority other than by complying with given 
commands or prohibitions. Contrary to all criteria of social stratifica- 
tion, authority does not permit the construction of a scale. So-called 
hierarchies of authority (as displayed, for example, in organization 
charts) are in fact hierarchies of the "plus-side" of authority, i.e., of 
the differentiation of domination; but there is, in every association, 
also a "minus-side" consisting of those who are subjected to authority 
rather than participate in its exercise. 

In two respects this analysis has to be specified, if not supple- 
mented. First, for the individual incumbent of roles, domination in 
one association does not necessarily involve domination in all others 
to which he belongs, and subjection, conversely, in one association does 
not mean subjection in all. The dichotomy of positions of authority 
holds for specific associations only. In a democratic state, there are 
both mere voters and incumbents of positions of authority such as 
cabinet ministers, representatives, and higher civil servants. But this 
does not mean that the "mere voter" cannot be incumbent of a position 
of authority in a different context, say, in an industrial enterprise ; con- 
versely, a cabinet minister may be, in his church, a mere member, i.e., 
subject to the authority of others. Although empirically a certain cor- 
relation of the authority positions of individuals in different associa- 
tions seems likely, it is by no means general and is in any case a matter 
of specific empirical conditions, it is at least possible, if not probable, 
that if individuals in a given society are ranked according to the sum 
total of their authority positions in all associations, the resulting pat- 
tern will not be a dichotomy but rather like scales of stratification ac- 
cording to income or prestige. For this reason it is necessary to empha- 
size that in the sociological analysis of group conflict the unit of analy- 
sis is always a specific association and the dichotomy of positions 
within it. 

As with respect to the set of roles associated with an individual, 
total societies, also, do not usually present an unambiguously dicho- 
tomic authority structure. There are a large number of imperatively 
coordinated associations in any given society. Within every one of 
them we can distinguish the aggregates of those who dominate and 
those who are subjected. But since domination in industry does not 
necessarily involve domination in the state, or a church, or other as- 
sociations, total societies can present the picture of a plurality of com- 

172 Social Structure, Group Interests , Conflict Groups 

peting dominant (and, conversely, subjected) aggregates. This, 
again, is a problem for the analysis of specific historical societies and 
must not be confounded with the clearer lines of differentiation within 
any one association. Within the latter, the distribution of authority 
always sums up to zero, i.e., there always is a division involving domi- 
nation and subjection. ^^ 

I need hardly emphasize that from the point of view of "settling" 
the concepts of power and authority, the preceding discussion has 
raised more problems than it has solved. I believe, however, that for 
the purposes of this study, and of a sociological theory of conflict, little 
needs to be added to what has been stated here. In order somewhat 
to substantiate this perhaps rather bold assertion, it seems useful to 
recapitulate briefly the heuristic purpose and logical status of the con- 
siderations of this section. 

I have introduced, as a structural determinant of conflict groups, 
the category of authority as exercised in imperatively coordinated as- 
sociations. While agreeing with Marx that source and level of income 
— even socioeconomic status — cannot usefully be conceived as deter- 
minants of conflict groups, I have added to this list of erroneous ap- 
proaches Marx's own in terms of property in the means of production. 
Authority is both a more general and a more significant social rela- 
tion. The former has been shown in our critique of Marx; the latter 
will have to be demonstrated by subsequent considerations and analy- 
ses. The concept of authority is used, in this context, in a specific sense. 
It is differentiated from power by what may roughly be referred to 
as the element of legitimacy j and it has to be understood throughout 
in the restricted sense of authority as distributed and exercised in im- 
peratively coordinated associations. While its "disruptive" or conflict- 
generating consequences are not the only aspect of authority, they are 
the one relevant in terms of the coercion model of society. Within the 
frame of reference of this model, ( i ) the distribution of authority in 
associations is the ultimate "cause" of the formation of conflict groups, 

^^ Inevitably, the qualifications introduced in the two preceding paragraphs are 
rather vague if stated merely in the abstract. They are, however, of the utmost im- 
portance for empirical analysis. By strictly postulating imperatively coordinated as- 
sociations as units of conflict analysis, we are able to consider, e.g., the relations be- 
tween industry and society as an empirical problem which allows of varying solutions 
in different historical contexts. Similarly we can, by this emphasis, regard subjection 
(and consequent deprivation) in several associations as a condition strengthening and 
intensifying conflict, but by no means necessary in historical situations. These and 
similar problems will become increasingly crucial as our investigation proceeds. 

Social Structure, Growp Interests , Conflict Groups 173 

and (2), being dichotomous, it is, in any given association, the cause 
of the formation of two, and only two, conflict groups. 

The first of these statements is logically an assumption, since it 
underlies scientific theories. It cannot as such be tested by observation j 
its validity is proven, rather, by its usefulness for purposes of explana- 
tion. We shall derive from this assumption certain more specific 
hypotheses which, if refuted, would take the assumption with them 
into the waste-paper basket of scientific theories. We assume in this 
sense that if we manage to identify the incumbents of positions of 
domination and subjection in any given association, we have identified 
the contenders of one significant type of conflicts — conflicts which oc- 
cur in this association at all times. 

As to the second statement, the one concerned with the dichotomy 
of authority positions in imperatively coordinated associations, it is not, 
I suggest, either an assumption or an empirical hypothesis, but an 
analytical statement. It follows from and is implicit in the very con- 
cept of authority that within specified contexts some have authority 
and others not. If either nobody or everybody had authority, the 
concept would lose its meaning. Authority implies both domination 
and subjection, and it therefore implies the existence of two distinct 
sets of positions or persons. This is not to say, of course, that there is 
no difference between those who have a great deal and those who have 
merely a little authority. Among the positions of domination there 
may be, and often is, considerable differentiation. But such differenti- 
ation, while important for empirical analysis, leaves unaffected the 
existence of a border line somewhere between those who have what- 
ever little authority and the "outs." Strictly speaking, an analytical 
statement which states that there is a dichotomy of authority positions 
is tautological j but as this example shows, there are tautologies which 
are worth stating. 

Having thus established the frame of reference and basic assump- 
tions of a sociological theory of conflict, we now turn to its more spe- 
cific elements — first with respect to the formation of conflict groups, 
then with respect to patterns of conflicts between these groups. 


The analytical process of conflict group formation can be described 
in terms of a model. Throughout, the categories employed in this 
model will be used in terms of the coercion theory of social structure. 
With this restriction in mind, the thesis that conflict groups are based 
on the dichotomous distribution of authority in imperatively coordi- 

1 74 Social Structure, Group Interests, Conflict Groups 

nated associations can be conceived of as the basic assumption of the 
model. To this assumption we now add the proposition that diflFeren- 
tially equipped authority positions in associations involve, for their 
incumbents, conflicting interests. The occupants of positions of domi- 
nation and the occupants of positions of subjection hold, by virtue of 
these positions, certain interests which are contradictory in substance 
and direction. In the case of incumbents of ruling positions, these in- 
terests, being "ruling interests" themselves, might also be described 
as values j however, in the present context I propose to retain the cate- 
gory of interest as a general term for the orientation of dominating 
and subjected aggregates. 

By postulating interests that are given and conditioned by posi- 
tions, we encounter once again a problem which we must now face 
squarely. In everyday language, the word "interest" signifies inten- 
tions or directions of behavior associated with individuals rather than 
with their positions. It is not the position, but the individual who "is 
interested in something," "has an interest in something," and "finds 
something interesting." It might indeed appear that the notion of in- 
terest is not meaningfully conceivable other than in relation to human 
individuals. Interests would seem to be psychological in the strictest 
sense. Yet the proposition of certain antagonistic interests conditioned 
by, even inherent in, social positions contains precisely this apparently 
meaningless assertion that there can be interests which are, so to say, 
impressed on the individual from outside without his participation. 

"As in private life we distinguish," says Marx, "between what a 
man thinks and says of himself and what he really is and does, so must 
we distinguish even more carefully in historical struggles the catch- 
words and fantasies of parties from their real organism and their 
real interests, their conception from their reality" (8, p. 38). Else- 
where, Marx equates a "common situation" with "common interests" 
(see 6, p. 187) and thereby shows that — as we do here — he bases his 
theory on a quasi-"objective," nonpsychological concept of interest. 
Geiger was the most extreme critic of such a concept. He regards it as 
"questionable whether one can speak at all of interest in an objective 
sense. Interest is above all something subjective . . ." (46, pp. 
127 f.). He believes that the (Marxian) postulate of interests ex- 
isting independently of individuals conceals a judgment "about the 
true advantage of others," for which "one would evidently have to 
possess an objective and universally valid scale of values" (p. 129). 
Since such a scale "plainly does not exist," it is not science that has 
guided Marx's pen here, but pure speculation. He imputes, according 

Social Structure y Grou-p Interests y Conflict Growps 175 

to Geiger, "true interests" to the proletariat which are independent of 
the wishes and goals of its members. But here "the proper analysis of 
the interest structure of social classes ends — religious mania alone 
speaks here" (p. 133). 

I would agree that the postulate of nonindividual class interests 
in its Marxian formulation requires criticism. But such criticism need 
not be directed against the possibility of a nonpsychological concept 
of interest. A concept of this kind responds, in fact, to a genuine need 
of sociological analysis. It is no accident that it appears and reappears 
frequently in the history of sociology — after Marx, in the works of 
Ratzenhofer, Small, Sumner, and many others. For L. Robbins, the 
distinction between "subjective" and "objective communities of in- 
terest" is the starting point of a critique of Marxj yet Robbins never 
doubts the usefulness of the category of "objective community of 
interest" (182, p. 112). M. Ginsberg refers to "aggregates . . . 
whose members have certain interests or modes of behavior in com- 
mon" without being "definite groups" (47, p. 40). Even Parsons not 
only frequently uses the clearly "objective" category of "vested in- 
terests," but also explicitly emphasizes the existence of common "ide- 
ologies" or "attitude systems" among those "who are structurally 
placed at notably different points in a differentiated social structure," 
i.e., occupy identical or similar social positions (67, p. 330). 

For purposes of the sociological analysis of conflict groups and 
group conflicts, it is necessary to assume certain structurally gen- 
erated orientations of the actions of incumbents of defined positions. 
By analogy to conscious ("subjective") orientations of action, it ap- 
pears justifiable to describe these as "interests," It has to be empha- 
sized, however, that by so doing no assumption is implied about the 
substance of these interests or the consciousness and articulate orienta- 
tion of the occupants of the positions in question." The assumption of 
"objective" interests associated with social positions has no psycho- 
logical implications or ramifications 5 it belongs to the level of socio- 
logical analysis proper. 

In saying that the Marxian notion of class interest requires criti- 
cism, I mean that Marx fuses, in this notion, generalization and spe- 
cific empirical observation in an impermissible manner. This becomes 
apparent if we look at the substance of socially structured interests. 
What is it that the occupants of positions of domination or subjection 

^^ This statement will be qualified below by the distinction of "latent" and 
"manifest interests." Stricdy speaking, it holds for latent interests only. 

1 76 Social Structurey Grouf Interests y Conflict Groups 

are "interested in" by virtue of their positions? Geiger is right in re- 
jecting Marx's attempt to answer this question in terms of material 
value conceptions. That "the realization of a socialist society" con- 
stitutes "the true interest of labor" is indeed an assertion for the (em- 
pirical) premises of which "the proof is missing" (46, pp. 130 f.). 
An assumption of this kind cannot be introduced by way of a postulate. 
The substance of socially structured "objective" interests can be de- 
scribed only in highly formal terms: they are interests in the main- 
tenance or modification of a status quo. Our model of conflict group 
formation involves the proposition that of the two aggregates of 
authority positions to be distinguished in every association, one — that 
of domination — is characterized by an interest in the maintenance of 
a social structure that for them conveys authority, whereas the other — 
that of subjection — involves an interest in changing a social condition 
that deprives its incumbents of authority. The two interests are in 

Max Weber has convincingly demonstrated that the problem of 
maintaining or changing given structures of authority can be ex- 
pressed, both conceptually and empirically, in terms of the basis of 
legitimacy of relations of authority. From our assumption of an at 
least latent conflict of interests in every imperatively coordinated as- 
sociation, it follows that the legitimacy of authority must always be 
precarious. There always is one aggregate of positions and their in- 
cumbents which represents the institutionalized doubt in the legitimacy 
of the status quo of the distribution of authority. In this sense, the 
proposition that there are "objective" interests in changing any given 
structure of authority might also be expressed in terms of the poten- 
tial illegitimacy of all relations of authority. Empirically, group con- 
flict is probably most easily accessible to analysis if it be understood as 
a conflict about the legitimacy of relations of authority. In every as- 
sociation, the interests of the ruling group are the values that constitute 
the ideology of the legitimacy of its rule, whereas the interests of the 
subjected group constitute a threat to this ideology and the social re- 
lations it covers. 

There are two further ways of elucidating the important notion of 
socially structured conflicts of interest. One of these has recourse to 
what is, if not a general "psychological law," at least an assumption im- 
plicit in many theories of economics, psychology, and sociology. In a 
word, this "law" might be described as the pleasure principle. It might 
be argued that the assumption of a fundamental human tendency to 
improve the balance of pleasure and pain, or gratification and depriva- 

Social Structure y Grouf Interests j Conflict Groups 177 

tion, implies inter alia that wherever there is authority those in posi- 
tions of domination would tend to defend their gratification, whereas 
those in positions of subjection are forced to attack existing conditions 
in order to remedy their deprivation/^ 

Plausible as this argument may sound, it is both unnecessary and 
somewhat misleading. It is unnecessary because the proposition that 
a conflict of interests is associated with authority positions in any as- 
sociation requires no recourse to more general assumptions. At the 
same time, the argument is misleading not only because (since Freud) 
we know that the pleasure principle is at best of restricted applicability, 
but above all because it might suggest that the "objective" interests of 
our model are after all psychological realities. Our proposition does 
not imply that the incumbents of positions equipped with "objective" 
interests will necessarily become conscious of these interests and act 
accordingly. While this may be probable, it is not as such required 
by the model of conflict group formation." 

I suggest, therefore, that a second line of argument is rather more 
appropriate for elucidating the notion of "objective" interests. This 
has the additional advantage of enabling us to abandon the awkward 
concept of "objective" interests in favor of more unambiguous and 
precise notions. In terms of the integration theory of society, social 
positions, with which we are here concerned, are significant, above all, 
as social roles. By roles are understood sets of role expectations, "pat- 
terned expectations defining the proper behavior of persons playing 
certain roles" (Parsons, 217, pp. 61 f.). "Proper" means, of course, 
within the frame of reference of integration theory, appropriate for 
the functioning of the social system and contributing to its integration. 
The notion of role expectations ascribes an orientation of behavior to 
social positions or roles. The individual "player" of roles may or 
may not internalize these role expectations and make them conscious 
orientations of action. If he does so, he is in terms of integration 
theory "adapted" or "adjusted" j if he does not do so, he is a "devi- 
ant." In any case, the assumption of certain "objective" expectations 
of behavior proves analytically useful. I suggest that the category of 

^^ In several earlier essays (cf, 42, p. 42 ; 206, p. 5 1 2 and note 8 1 ) , I have in fact 
argued along these lines. However, today I would regard this argument, if not as 
false, then as misleading, and prefer the second approach indicated below. 

^*Thus, the model does not prejudice the (empirical) problem of whether in 
any given case those in subjection want to attack those in domination or not. Con- 
scious motivations are, from the point of view of conflict theory, largely a matter of 
empirical research. 

178 Social Structure y Grouf Inter ests^ Conflict Groups 

interest in the coercion theory of society must be understood in strict 
analogy to that of role expectation. The "objective" interests under 
discussion are in fact role interests, i.e., expected orientations of be- 
havior associated with authority roles in imperatively coordinated as- 
sociations. Again, the individual incumbent of roles may or may not 
internalize these expectations. But in our context he behaves in an 
"adapted" or "adjusted" manner if he contributes to the conflict of 
contradictory interests rather than to the integration of a social system. 
The individual who assumes a position in an association finds these role 
interests with his position, just as he finds certain role expectations 
from the point of view of the social system. For diflFerent purposes of 
sociological analysis, different aspects of its basic unit — the position- 
role — are relevant j roles, too, have two faces. In our context they 
figure primarily as sets of expected interests within imperatively co- 
ordinated associations.^^ 

For certain purposes of the theory of conflict group formation, it 
will prove useful to replace the concept of role interests by another one 
which makes its relation to the incumbents of authority positions even 
more apparent. Role interests are, from the point of view of the 
"player" of roles, latent interests y i.e., undercurrents of his behavior 
which are predetermined for him for the duration of his incumbency 
of a role, and which are independent of his conscious orientations. As 
such they can, under conditions to be specified presently, become con- 
scious goals which we shall correspondingly call manifest interests^^ 
By contrast to latent interests, manifest interests are psychological 
realities. They describe — as Geiger demands for all interests — "the 
fact that emotion, will, and desire of a person are directed toward 
some goal" (although we presuppose that this goal is "some goal" 
only in a substantial, not in a formal, sense) . The specific substance of 
manifest interests can be determined only in the context of given 
social conditions j but they always constitute a formulation of the is- 
sues of structurally generated group conflicts of the type in question. 
In this sense, manifest interests are the program of organized groups. 

^^ The frultfulness of this necessarily rather abstract analysis can be demonstrated 
only by empirical analysis. For some indication of the potentialities of the position 
here suggested, cf. the discussion of codetermination in industry in Chapter VII. 

^° These two terms which — apart from their evident meaning as words — refer 
back to Merton's distinction of "manifest" and "latent functions" and further to 
Freud's categories of "manifest" and "latent dream contents," were first proposed by 
me in an essay on class (42, pp. 11 f.). Some of the purely conceptual problems are 
discussed at somewhat greater length in that essay. 

Social Structure y Group Interests y Conflict Groups 179 

The psychological formations which we have called "manifest 
interests" are evidently similar to what in Marxian, and more gen- 
erally in sociological, literature is usually referred to as "class con- 
sciousness." However, it is necessary to distinguish strictly the philo- 
sophical and speculative elements of this concept as used, for example, 
by Marx and Lukacs (21), from the observable articulate interests of 
organized groups. Class consciousness in the sense of manifest in- 
terests is a "real category." Its existence and substance can in principle 
be discovered by interviews. The conception of a "false conscious- 
ness" can, in the context of the categories here employed, be meaning- 
ful only for manifest interests which are not adapted to the latent 
interests underlying them j even then, it is a highly problematical con- 
ception. In terms of a scientific theory which is supposed to explain 
problems of reality, the statement that a large group of people thinks 
"falsely" is plainly meaningless. "One has to be a philosopher of real 
dialectics to be serious about such nonsense" (Geiger, 46, p. 114). 
While latent interests are, in a psychological sense, "nonexistent," 
manifest interests are always realities in the heads of the occupants 
of positions of domination or subjection in associations. It is the task 
of the theory of class formation to establish a systematic connection 
between the two categories of latent and manifest interests. 


So far our discussion has left undecided the question as to what 
kind of aggregates conflict groups are. We have established the de- 
terminant of group conflicts in the sense of this study; we have also 
established the categories characteristic of conflict groups on the nor- 
mative level of analysis; but conflict groups are evidently not norma- 
tive phenomena but real groupings which as such form part of the 
substratum of society. In describing these groupings, the categories 
of quasi-group and interest group are essential. 

We have postulated two conflicting orientations of latent interests 
as characteristic of the role structure of imperatively coordinated as- 
sociations. By implication, this means, of course, that the authority 
positions equipped with expected interests as well as their incumbents 
have at least one attribute in common. In a significant sense, the oc- 
cupants of identical authority positions, i.e., either of positions of 
domination or of positions of subjection, find themselves in a common 
situation. Being united by a common, potentially permanent, char- 
acteristic, they are more than mere masses or incoherent quantities. 

i8o Social Structure y Group Interests y Conßict Groups 

At the same time, the incumbents of like authority positions in an 
association do not in any sociologically tenable sense constitute a group. 
Just as all doctors, or all inhabitants of Berlin, do not as such constitute 
social groups, the occupants of positions with identical latent interests 
are not a group. For groups, a feeling of belongingness is as constitu- 
tive as a minimum of organization j but both are explicitly not de- 
manded by the concept of latent interests. The aggregates of incum- 
bents of positions with identical role interests are at best a potential 
group. Following M. Ginsberg we shall use for this particular type 
of social grouping the term quasi-grouf. "Not all collectivities or 
aggregates form groups. Groups are masses of people in regular con- 
tact or communication, and possessing a recognizable structure. There 
are other aggregates or portions of the community which have no 
recognizable structure, but whose members have certain interests or 
modes of behavior in common, which may at any time lead them to 
form themselves into definite groups. To this category of quasi- 
groups belong such entities as social classes, which, without being 
groups, are a recruiting field for groups, and whose members have 
certain characteristic modes of behavior in common" (47, p. 40). 

A note of caution is required with respect to the "modes of be- 
havior" included by Ginsberg in his definition of quasi-groups. The 
constituent element of the type of quasi-groups with which we are 
here concerned is the community of certain latent interests. Latent 
interests are not psychological phenomena; quasi-groups based on 
them might therefore be called a mere theoretical construction. They 
are "theoretical phenomena," i.e., units constructed for the purpose of 
explaining problems of social conflict. They are, as Ginsberg plausibly 
states, "recruiting fields for groups." For purposes of a sociological 
theory of group conflict it is useful to reduce the actual conflict groups 
of empirical associations to larger aggregates which form part of their 
structure and consist of the incumbents of roles endowed with like ex- 
pectations of interest. Only by doubtful analogy can we speak of 
"members" of such aggregates or quasi-groups. Thus it transcends 
the legitimate possibilities of theory construction to postulate common 
modes of behavior for these "members." 

On the other hand, common modes of behavior are characteristic 
of interest groups recruited from larger quasi-groups. Interest 
groups are groups in the strict sense of the sociological term; and they 
are the real agents of group conflict. They have a structure, a form of 
organization, a program or goal, and a personnel of members. If 
Ginsberg demands for such groups "regular contact or communica- 

Social Structure y Group Interests y Conßict Groups 1 8 1 

tion," however, this applies only in an indirect sense. Interest groups 
are always "secondary groups" j their members are in contact with 
each other only by virtue of their membership or by way of their 
elected or appointed representatives. One might emphasize the dif- 
ference between interest groups and primary groupings such as family 
or friendship by calling them with Maclver "associations" or with 
Malinowski "institutions." However, it seems to me that the concept 
of interest group is sufficiently unambiguous if, apart from termino- 
logical considerations, we keep the modern political party in mind as 
an example of such organizations. 

It is perhaps necessary to emphasize that the groupings with which 
we are here concerned are by no means satisfactorily described by the 
concepts of quasi-group and interest group. This is especially evident 
in the case of interest groups. The statement that conflict groups are 
interest groups is meaningful, but incomplete. The category of in- 
terest group is a general category j virtually any secondary group can 
be regarded as an interest group — a chess club as well as an occupa- 
tional association, a football team as well as a political party or a trade 
union. The specific difference of the quasi-groups and interest groups 
with which we are concerned in this study accrues from their origin in 
the authority structure of associations or, to put it differently, from the 
formal characteristic of their underlying (latent or manifest) interests 
as interests related to the legitimacy of relations of domination and 
subjection. This limitation clearly excludes the chess club, the foot- 
ball team, and the occupational association, while it leaves us to con- 
sider groupings such as trade unions and political parties. Whenever 
we refer in the subsequent analysis to quasi-groups and interest groups 
without specifically stating this limitation, this is merely an abbrevi- 
ated way of referring to conflict groups as they emerge from the 
authority structure of associations. 

The empirical problem of the genesis of interest groups consti- 
tutes the subject matter of the following section of this chapter. 
On a more formal level, however, the problem of the relation be- 
tween quasi-groups and interest groups may be raised here. In what 
sense are interest groups, such as political parties, to be regarded as 
representative of the quasi-groups that can be inferred behind them? 
Can the same quasi-group become a recruiting field for several in- 
terest groups.? In principle, the possibility intimated by the latter 
question has to be answered in the affirmative. From the point of 
view of conflict theory, competing trade unions of, say. Christian 
and Socialist description originate from the same quasi-group. Em- 

1 82 Social Structure y Group Interests y Conflict Groups 

pirically, interest groups are always smaller than their recruiting 
fields, the quasi-groups. They are subsets of the sets constituted by 
quasi-groupsj and the identity of set and subset remains a limiting 
case. One might compare the relation of the two with that of the 
members and the voters of one political party. Furthermore, a num- 
ber of specific intervening variables may disturb the immediacy of 
the relation between given quasi-groups and interest groups. While 
quasi-groups, being in the nature of a theoretical construction, are 
unequivocally defined, organized interest groups may supplement 
the interests accruing from authority structures by a multitude of 
other and independent goals and orientations. This is merely another 
expression for the fact that interest groups are "real phenomena," 
and that, like all such phenomena, they cannot be completely de- 
scribed by one attribute. Thus, the theory of group conflict involves 
no statement about the empirical variety of interest groups. It con- 
centrates on one of their aspects: on their function in social conflicts 
as units of manifest interests which can be explained in terms of latent 
role interests and their aggregation in quasi-groups." 


"It is a matter of no small interest," Ginsberg adds to his defi- 
nition of quasi-groups, "to determine at what point these looser con- 
figurations crystallize into associations" (47, p. 41). The categories 
of quasi-group and interest group mark the two foci of the analysis 
of conflict group formation, but they do not describe the connecting 
lines between them. It will now be our task to examine the condi- 
tions under which a "class in itself" becomes a "class for itself." 
Perhaps the negative side of this problem is of even greater impor- 
tance. We shall want to ascertain the conditions under which the 
organization of interest groups does not take place despite the pres- 
ence of quasi-groups of latent interests in an imperatively coordi- 
nated association. This is evidently a matter of ascertaining possible 
intervening variables which we shall comprehend under the collec- 
tive term of "structural conditions of organization," 

In dealing with the empirical process of development of classes 

^^ To illustrate this rather abstract formulation: for the theory of conflict, so- 
cialist parties are of interest not as instruments of workers' education or as clublike as- 
sociations, but merely as forces in social conflicts. The same party may function in 
many ways other than as an interest group, but only the latter aspect is in question in 
the present analysis. 

Social Structure y Group Interests y Conflict Groups 183 

Marx has touched upon this problem at many points. Among these 
there is one which is particularly illuminating for our present context. 
At the end of his essay on the 1 8th of Brumaire of Louis Bonaparte, 
Marx is dealing with "the most numerous class of French society, 
the small independent peasants" (8, p. 104). Marx states, to begin 
with, that these peasants, by virtue of their situation, their conditions 
of existence, their way of life, and their (latent) interests, constitute 
a "class," namely, a quasi-group. One would therefore expect a 
political organization or interest group to grow out of their midst. 
However, precisely this did not happen. In so far as the identity of 
the (latent) interests of the peasants "does not produce a community, 
national association and political organization, they do not constitute 
a class" (p. 105). In explaining this surprising fact, Marx refers to 
conditions of the kind of the intervening variables in question here: 
"The small independent peasants constitute an enormous mass, the 
members of which live in the same situation but do not enter into 
manifold relations with each other. Their mode of production iso- 
lates them from each other instead of bringing them into mutual 
intercourse. This isolation is strengthened by the bad state of French 
means of communication and by the poverty of the peasants. . . . 
Every single peasant family is almost self-sufficient . . . and thus 
gains its material of life more in exchange with nature than in inter- 
course with society" (p. 104). The brilliant conclusion Marx draws 
from this analysis — namely, that Louis Bonaparte is trying to justify 
his claim for power by reference to this quasi-group of peasants 
whose interests are condemned to latency — will concern us less here 
than the problem impressively demonstrated by it. Under certain 
conditions, quasi-groups may persist as such without interest groups 
emerging from them. What are these conditions, and under which 
conditions do interest groups come to be formed? 

It may be useful to begin by clarifying the logical status of a 
generalizing answer to this question. The categories of latent and 
manifest interest, quasi-group and interest group, constitute the ele- 
ments of a model of conflict group formation. Under ideal condi- 
tions, i.e., if no variables not contained in this model intervene, the 
analytical^* process of conflict group formation can be represented 
as follows. In every imperatively coordinated association, two quasi- 

^^ This needs to be emphasized. We are not here concerned with the chronologi- 
cal development of conflict groups. Marx commits at this point an error of hypostasis. 
By asserting the analytical sequence as a chronological one, he transforms the "theoreti- 
cal phenomenon" quasi-group into a "real phenomenon" in an impermissible manner. 

1 84 Social Structurey Growp Interests^ Conflict Groups 

groups united by common latent interests can be distinguished. Their 
orientations of interest are determined by possession of or exclusion 
from authority. From these quasi-groups, interest groups are re- 
cruited, the articulate programs of which defend or attack the legiti- 
macy of existing authority structures. In any given association, two 
such groupings are in conflict. This model of conflict group forma- 
tion is as such complete and suffices for all purposes of theoretical 
analysis. In principle, little need be added to it, and what additions 
are required are in the nature of refinements. However, for purposes 
of empirical analysis the model may — as the example of Marx's 
"i8th of Brumaire" shows — be useful as a guide to relevant prob- 
lems, but it is as such incomplete. As soon as we pass from the level 
of model construction to that of the explanation of empirical prob- 
lems, the premise of ideal conditions is no longer given. We en- 
counter intervening variables the identification of which is our con- 
cern at this point. 

Determination of the structural conditions of organization re- 
quires the formulation of empirical generalizations. Here, the se- 
curity of theoretical postulates and constructions gives way to the 
precarious attempt to expose general facts of social life, to classify 
them, and to formulate them hypothetically. It might be argued 
that this step transcends the limits of theory and need not therefore 
be taken here. It passes, indeed, beyond the limits of model con- 
struction. Nevertheless, it has to be taken in order for the model 
to be freed of the suspicion of analytical uselessness. The attempt 
to classify intervening variables in a generalizing manner is all the 
more important, since without it ad hoc hypotheses and additions have 
free reign." For the sake of the empirical applicability of the model 
proposed here it is necessary to indicate in general the structural 
conditions of the organization of conflict groups. Completeness is 
intended but of course not by any means guaranteed. 

Malinowski endeavored, in his attempt to classify the character- 
istics of what he calls "institutions" (and what might equally well 
be called "associations" or simply "organized groups"), to render, 
as he puts it, this category "more serviceable in field-work" (212, 
pp. 52 ff.). Malinowski lists at this point six features which are of 
importance also for interest groups of the kind discussed here: such 
groups require a charter, a personnel, certain norms, a material in- 

^^ In this sense, Marx's explanation in the "18th Brumaire" is of course also an 
ai hoc addition, if a brilliant one. All subsequent considerations in this chapter are on 
the level of supplementing a theoretical model by empirical generalizations. 

Social Structure y Group Interests, Conflict Groups 185 

strumentarium, certain regular activities, and an "objective" func- 
tion. The latter characteristic is in our case part of the underlying 
model. Otherwise, however, Malinowski's list contains an important 
sector of the empirical conditions of organization of interest groups. 
I shall call these conditions the technical conditions of organization. 
Without a charter, certain norms, a personnel, and certain material 
requisites, interest groups cannot be formed, even if it is justified 
to assume that quasi-groups exist. Obvious as these conditions may 
appear, at least two of them — "charter" and "personnel" — are of 
considerable significance, as will become apparent if we specify them 
a little. 

It is a commonplace that groups cannot exist without members 
and, in that sense, without a personnel. Moreover, since we have 
postulated the presence of a personnel in the quasi-groups from which 
interest groups emerge, this condition does not at first sight appear 
to be a genuine intervening variable. It is, indeed, not the total mem- 
bership of an interest group which is in question here, but that sector 
of the membership which can be described as the leading group or 
cadre. For an organized interest group to emerge from a quasi-group, 
there have to be certain persons who make this organization their 
business, who carry it out practically and take the lead. Every party 
needs its founders. The availability of founders in this sense, how- 
ever, is by no means given in our model nor can it be. It is an addi- 
tional empirical condition of conflict group formation. As such it is 
a necessary, although not sufficient, condition of organization. To 
stipulate a leading group as a prerequisite of the organization of in- 
terest groups must not be misunderstood to mean that conflict groups 
are based on the goals and actions of a handful of leaders. The avail- 
ability of possible organizers, founders, and leaders is essentially a 
technical prerequisite which must be satisfied for unorganized quasi- 
groups to be transformed into organized interest groups. The or- 
ganizers are one of the ferments, not the starting point or cause of 
organization. That without them organization is impossible has been 
demonstrated convincingly — in so far as it is not self-evident — by 
Marx at the place in the "i8th Brumaire" quoted above, and above 
all in the "Communist Manifesto." 

Marx has realized, also, that the creation of a charter is not an 
automatic process. Malinowski defines the charter of an organiza- 
tion as the "system of values for the pursuit of which human beings 
organize" (212, p. 52). In the particular case of conflict groups these 
values consist of what we have called "manifest interests." While 

1 86 Social Structure y Growp Interests y Conflict Groups 

latent interests are nonpsychological orientations implicit in the social 
structure of roles and positions, manifest interests are articulate, for- 
mulated (or at least formulable) programs. They entail specific 
claims related to given structures of authority. The articulation and 
codification of such interests is again a process that presupposes certain 
conditions. Either there must be a person or circle of persons who 
take on themselves the task of articulation and codification, or, al- 
ternatively, an "ideology," a system of ideas, must be available which 
in a given case is capable of serving as a program or charter of groups. 
As evidence for the first, it seems sufficient to refer to the role of 
the political ideologist Marx for the organization of the socialist 
movement j as evidence of the latter alternative, to the role of a 
certain interpretation of Calvinism for early English capitalists. 
Ideologies understood as articulated and codified manifest interests 
are again but a technical condition of organization. Ideologies do not 
create conflict groups or cause conflict groups to emerge. Yet they 
are indispensable as obstetricians of conflict groups, and in this sense 
as an intervening variable. 

Even if we are given not only quasi-groups with common latent 
interests, but leaders and ideologies as well — if, in other words, the 
technical conditions of organization are present — it is still not justi- 
fied to make the empirical inference that interest groups will be 
formed. A second category of prerequisites which have to be satisfied 
for organization to be possible will be described here as the political 
conditions of organization. The totalitarian state is probably the most 
unambiguous illustration of a social situation in which these condi- 
tions are not fulfilled, and in which therefore at least oppositional 
interest groups cannot emerge despite the presence of quasi-groups 
and latent interests.'" Where a plurality of conflicting parties is not 
permitted and their emergence suppressed by the absence of freedom 
of coalition and by police force, conflict groups cannot organize them- 
selves even if all other conditions of their organization are present. 
The study of the possibilities and actual types of group conflict under 
such conditions is a problem of sociological analysis of the highest 

^° Technically similar conditions obtain in many preindustrial societies. In terms 
of the political conditions of organization the restriction of this study to industrial so- 
cieties can be well illustrated. In all preindustrial societies group conflict is seriously 
impeded by the absence of certain political conditions (the political "citizenship 
rights"). It would be a matter for separate analysis to investigate forms of group con- 
flict in these societies. 

Social Striicturey Growp Interests., Conflict Growps 187 

importance. There is a starting point, here, of the analysis not only 
of "underground movements" and the development of revolutions, 
but more generally of structure and dynamics of totalitarian states. 
But this type of problem can be merely intimated here, since we are 
for the time being concerned with formulating the general structural 
conditions of organization. We can maintain that the political per- 
missibility of organization is one of the additional intervening pre- 
requisites of conflict group formation. 

Apart from technical and political conditions, some, in the narrow 
sense, sodal conditions of organization are of importance for the for- 
mation of interest groups. Among these we find the condition of 
communication between the "members" of quasi-groups emphasized 
by Marx in the case of French peasants. If an aggregate within an 
association can be described as a community of latent interests, is also 
provided with the technical and political possibilities of organization, 
but is so scattered topologically or ecologically that a regular con- 
nection among the members of the aggregate does not exist and can 
be established only with great difficulty, then the formation of an 
organized interest group is empirically most unlikely. However, 
important as this premise of organization is, the generalization seems 
tenable that its significance is steadily diminishing in industrial so- 
cieties with a highly developed system of means of communication. 
In advanced industrial societies this condition may be assumed to be 
generally given 5 it enters, therefore, into the analysis of conflict group 
formation as a constant. 

This is not the case, however, with another social condition of 
organization the implications of which will occupy us a good deal 
more. Empirically, the formation of organized interest groups is 
possible only if recruitment to quasi-groups follows a structural pat- 
tern rather than chance. By this condition, the group described by 
Marx as lumpenfroletariat is excluded from conflict group forma- 
tion.^^ Persons who attain positions relevant for conflict analysis not 
by the normal process of the allocation of social positions in a social 
structure, but by peculiar, structurally random personal circumstances, 
appear generally unsuited for the organization of conflict groups. 
Thus the lowest stratum of industrial societies is frequently recruited 
in manifold but structurally irrelevant ways: by delinquency, ex- 

^^ In my study of unskilled industrial workers in England (143) I have tried to 
show in detail how this one condition is capable of excluding a large group of people 
from political activity. 

1 8 8 Social Structure , Group Interests ^ Confiict Groups 

treme lack of talent, personal mishaps, physical or psychological in- 
stability, etc. In this case, the condition of structural recruitment is 
not satisfied, and conflict group formation cannot be expected. 

From the empirical conditions of the organization of conflict 
groups thus briefly sketched, we can, by way of generalization, derive 
a number of social constellations which are unfavorable if not pro- 
hibitive for conflict group formation and group conflict. Here, again, 
I shall confine myself to giving an indication. One constellation re- 
sisting conflict group formation, namely, that of the totalitarian state, 
is directly given in the formulated conditions. A second important 
constellation can be d-efined by combining several of the factors men- 
tioned. If imperatively coordinated associations are either themselves 
just emerging or subject to radical change, the probability is small 
that the quasi-groups derived from their authority structure will lead 
to coherent forms of organization. Examples for this may be seen 
in the early stages of industrial development, or in societies im- 
mediately after social revolutions (such as the Soviet Union in the 
1920's)." In both cases authority structures, latent interests, and 
quasi-groups are present. But in both cases it seems reasonable to 
assume that the absence of leaders and ideologies as well as the still 
unpatterned and unnormalized recruitment to the relevant positions 
stand in the way of conflict group formation. In this sense, it seems 
feasible to attempt to reformulate Marx's problem of the gradual 
formation of classes in the course of industrialization. 

The empirical conditions of organization have been described here 
as prerequisites of conflict group formation. However, their effect 
goes beyond the process of emergence of conflict groups. These fac- 
tors are relevant, also, as variables affecting organized interest groups. 
They must then be understood, of course, as continua which permit 
gradations. A relative lack of technical, political, and social condi- 
tions of organization can hamper organized interest groups in their 
operation, and it can, indeed — which is apparent in the case of the 
political conditions — result in their disintegration. Ideologies may 
lose their value as programs and their validity, especially if signifi- 
cant structure changes have occurred since their formulation. Parties 
may go through a "leaderless" period. Modes of recruitment to 
quasi-groups may change. Some of the problems of this type we 
shall encounter again in the empirical analyses of the final chapters 

^^ Here, the absence of the political conditions of organization is of course an 
added obstacle to interest group formation. 

Social Structure y Growp Interests y Confict Groups 189 

of this study. First of all, I propose to look a little more closely into 
the social and psychological characteristics of the conflict groups stipu- 
lated by our model. 


Conflict groups are certainly also psychological phenomena. We 
have defined organized interest groups by manifest interests, i.e., a 
characteristic of clearly psychological reality. Beyond that it is prob- 
able that the formation of interest groups presupposes certain psy- 
chological conditions apart from its technical, political, and social 
prerequisites. For example, identification with the expectations asso- 
ciated with authority roles is a condition of conflict group formation 
which has been mentioned already. It seems questionable, however, 
whether we comprehend this or any other problem of the psychology 
of conflict groups any better by following Warner (75, 100) and 
many other predominantly American social scientists in restricting 
our questions to an investigation of how people rank themselves or 
each other in society. From the point of view of a theory of social 
conflict, an approach of this kind is meaningless because it seeks to 
establish the notion of conflict groups empirically by deriving it from 
the opinions of a (more or less) representative sample of people. 
Although investigations concerned with people's self-evaluation and 
ranking of others are not entirely without interest for the sociologist, 
they are useless for our problem. They substitute for the effort of 
a theoretical derivation of significant problems of research the skills 
of drawing up questionnaires and interpreting interview findings. 
This is not to say, of course, that there are no social psychological 
studies at all which might be helpful for conflict theory. Four inde- 
pendent yet substantially similar studies — those of Centers (38), 
Hoggart (52), Popitz (69), and Willener (76) — are of consider- 
able significance here, and will occupy us in some detail as we come 
to analyze post-capitalist society in terms of conflict theory. At this 
point, however, we are concerned not with empirical findings but 
with the type of problem suggested by conflict group formation on 
the level of psychological research. A note on this may facilitate our 
further considerations. 

We can maintain, to begin with, that the model of group forma- 
tion formulated in this chapter does not presuppose or imply any 
psychological assumption. In this model, individual behavior figures 
— as indeed it should in sociological analysis — as a constant. How- 
ever, in applying this model to specific empirical conditions, we in- 

190 Social Structure y Growp Interests ^ Conflict Groups 

evitably encounter factors and variables of a psychological as well as 
a sociological nature. The attempt to present a general formulation 
of the three most important types of problems which come up here 
cannot of course relieve us of testing the validity of such generali- 
zations anew in every specific instance. 

(i) In exploring the emergence of interest groups from quasi- 
groups we encounter a problem which is logically equivalent to that 
of "deviant behavior" in the integration theory of social structure. 
Our model postulates quasi-groups and interest groups on the basis of 
the position of roles in imperatively coordinated associations. Thus, 
his "class situation" is forced upon the individual with the position 
he assumes in an association. On the level of interest groups, conscious 
and intentional participation of individuals is moreover assumed by 
definition. Manifest interests are psychological realities. Evidently, 
their presence can be neither simply assumed nor inferred from the 
presence of the technical, political, and social conditions of organi- 
zation. Rather, we have to ask under which psychological conditions 
individuals may orient themselves in accordance with or in contra- 
diction to the expected interests of their position. The worker who 
behaves as if he is not in a position of subjection is like the entrepre- 
neur who acts as if he is not in a position of domination, a "deviant" 
within the association of industry whose behavior requires a psycho- 
logical explanation. I shall refrain from offering vague suggestions 
about possible causes of such behavior j evidently this is a matter of 
social psychological research. ^^ It may be added that if any patterns 
of deviance from postulated norms or expectations in this sphere of 
social action are discovered, these patterns would by implication con- 
stitute the psychological conditions of organization which would thus 
supplement the three types of conditions distinguished in the pre- 
ceding section. 

(2) A second aspect of the psychology of conflict groups would 
seem to be of a phenomenological character. It is worth exploring 
the psychic determinants and features of manifest interests and of 
the solidarity of interest groups based on these. The sociological 
aspect of manifest interests is defined by the model of conflict group 
formation; moreover, this model permits the discovery of the sub- 
stance of manifest interests in given social situations. However, it is 
a matter of psychological research to examine in given cases whether 

^^ The theory of reference group behavior might here be helpful. It could ac- 
count for the weight of conflict groupings in the minds and actions of individuals by 
comparison to other types of relation. 

Social Structure J Group Interests, Conflict Groups 191 

specific manifest interests correlate with certain types or character- 
istics of personality, and to what extent role interests mold and pene- 
trate the personalities of their carriers.^* For this purpose it would 
be desirable to have an operational definition of manifest interests 
which establishes, in general, indices and methods of measuring the 
presence as well as the intensity of manifest interests. 

(3) The hypothesis seems plausible that weight and intensity of 
manifest group interests within the individual personality decrease 
as social mobility and the openness of conflict groups increase. The 
easier it is for the individual to leave his conflict group, the less likely 
is he to engage his whole personality in group conflicts and the more 
marginal is his authority role likely to remain. But the degree of 
openness of conflict groups is probably not the only determinant of 
the intensity of identification with conflict groups. According to our 
theory, an individual can belong to several conflict groups simulta- 
neously, in so far as he may play roles in several associations (e.g., 
industry and political society) simultaneously. It appears that the 
factual weight of an individual's belongingness to diff^erent associa- 
tions within the ensemble of his social personality also influences the 
intensity of his solidarity with any one conflict group to which he 

The problem indicated by these hypotheses is at least in part 
psychological. It is related to the problem of "class culture" which 
frequently appears in sociological literature. "One of the significant 
properties ... of the phenomenon of class consists in the fact that 
the members of a class behave towards each other in a way charac- 
teristically different from that towards members of other classes, that 
they stand in a closer relation to each other, understand each other 
better, cooperate more easily, join together and close themselves off 
from the outside, look at the same sector of the world with equally 
disposed eyes and from like points of view" (Schumpeter, 27, p. 
152). "Subjectively regarded, class differences rest upon the de- 
velopment of sentiments or groups of emotional dispositions. These 
are of three sorts. There is, first, a feeling of equality in relation to 
members of one's own class, a feeling of being at ease with them, a 
consciousness that one's mode of behavior will harmonize with the 
behavior of the others. There is, secondly, a feeling of inferiority 
to those above in the social hierarchy, and, thirdly, a feeling of su- 

^* This is rather a delicate task, since it is necessary to avoid the error of correlating 
group belongingness and personality type in a linear fashion, or even by way of as- 
sumption. See the discussion of Mayo's work in Chapter III above. 

192 Social Structure y Growp Interests ^ Conßict Groups 

periority to those below" (Ginsberg, 47, pp. 160 f.). "Social class 
is a derivative of the whole social personality of the individual, not 
of a mere facet of it, such as some technical equipment and the in- 
terests it may create. Social class is a human aggregation which has 
not been submitted to that splitting of individuality into its associative 
elements so subtly analyzed by Simmel. Each member mirrors in 
the microcosm of his personality the many featured image of his 
class" (Marshall, 58, p. 100). These assertions and assumptions — 
which are representative for a large number of authors — had to be 
quoted here in some detail, because they constitute a potential limi- 
tation of conflict theory to particular historical situations. They raise 
once again the problem of the feasibility of the concept of class. 
While postponing this aspect for a little while, we must emphasize 
here that conflict groups in the sense of this study differ radically 
from classes as described by Schumpeter, Ginsberg, and Marshall in 
the passages quoted. 

Conflict groups in the sense of our model describe primarily 
groupings based on positions in imperatively coordinated associations. 
Just as, from the point of view of the individual, his position in an 
association constitutes but one potentially very small sector of his 
social personality, his membership in a conflict group refers merely 
to a small segment of his personality. His conflict group claims — to 
use an illustration — the entrepreneur as an occupant of a position of 
domination in the enterprise, while leaving his behavior as husband, 
father, member of a church, a club, even as voter in principle un- 
decided. If there is a connection between the behavior of an indi- 
vidual by virtue of his membership in a conflict group and his total 
social behavior, this signifies a special case and cannot be postulated 
in general, just as the connection between industry and society is not 
determined a priori. The connection or correlation between conflict 
behavior and total social behavior is empirically variable and a matter 
for specific investigation. We have rejected Marx's and Burnham's 
alleged postulate of an identity of the authority structures of industry 
and the state, because it has revealed itself as a speculative dogma. 
Analogously, we must also reject the general correlation asserted by 
Schumpeter, Marshall, and Ginsberg between conflict group mem- 
bership and social personality^ again, this is no more than a generali- 
zation derived from a single observation." If we take this step, we 

^^ It may be suspected that all three authors are talking of social strata rather than 
conflict groups; and for strata their criteria may indeed hold. But the effort of the 
distinction between the analysis of social stratification and that of conflict must be sus- 
tained on all levels. 

Social Structure y Group Interests, Conflict Groups 193 

have reduced the problem of "class culture" (i.e., of the connection 
between membership in conflict groups and general social behavior) 
to a problem of empirical psychology. A continuous scale might be 
constructed, ranging from complete identity (a correlation of +1) 
of conflict behavior and general social behavior to complete irrele- 
vance (a correlation of o) of conflict behavior for other social be- 
havior. At a later stage, the importance of this step for the analysis 
of conflict in advanced industrial societies will become apparent. The 
authority role of the individual and such patterns of behavior as can 
be inferred from it are an independent variable the connection of 
which with other aspects of social behavior is theoretically indeter- 
minate and can be established only by empirical observation. Whether 
the whole personality of the individual is molded by his belonging 
to a conflict group, or whether the individual acts as a member of a 
conflict group only for limited periods of time (such as during his 
working time) and in limited social relations (such as in his capacity 
as trade unionist) while being guided by entirely different norms at 
other times and in other relations — this is a problem the solution of 
which suggests types of conflict groups and degrees of intensity of 
conflicts but has nothing to do with the existence of such groups and 
conflicts. At least in part, this solution is evidently a task of psycho- 
logical research. 

"elites" and "ruling classes" 

Our model of conflict group formation stipulates the existence of 
two opposed groupings in any given association. Each of these groups 
shares certain features, and each differs from the other by contra- 
dictory orientations of interest. Before concluding the abstract dis- 
cussion of the model and the examination of some of its empirical 
consequences we may ask what, if anything, can be stated in general 
about the two groups thus distinguished. Independent of particular 
empirical conditions, are there any features that characterize or other- 
wise distinguish the occupants of positions of domination and their 
interest groups from those of positions of subjection? It appears 
useful to discuss this problem with reference to the theories of three 
sociologists whose work is here representative and has heretofore in 
this discussion deliberately been mentioned only occasionally. I mean 
Pareto (25), Mosca (24), and Aron (34), whose conceptions re- 
semble ours in several points. Of the three, Mosca takes the most 
explicit stand on the problem at hand, and his conception will there- 
fore require particular attention. 

The chief element of the model of class formation consists in the 

1 94 Social Structure y Group Interests, Conflict Groups 

explanation of conflicts of interest groups in terms of quasi-groups 
determined by the distribution of authority in imperatively coordi- 
nated associations. We share this emphasis on authority structures 
with all three authors mentioned, whose work might therefore be 
described as the proximate origin of a theory of conflict of the type 
here proposed.^® Since they argue in terms of authority, Pareto, 
Mosca, and Aron also operate with a two-class model. It is charac- 
teristic of all of them, however, that they concentrate their attention 
— unlike Marx, Weber, and many others — on the group possessing 
authority, the members of which occupy, in other words, positions of 
domination. We shall presently consider some of the implications 
of this emphasis on dominating groups for the analysis of subjected 
groups and of group conflict in general. In describing dominating 
conflict groups the authors in question use primarily two concepts. 
Mosca refers almost exclusively to the "political class" which, in the 
German and English translations of his Elementi di Scienza Political 
has become a "ruling class." Pareto introduces for this group the 
much-disputed category of "elite" j however, he distinguishes "gov- 
erning" and "nongoverning" elites (25, p. 222) and devotes as much 
attention to the latter as to the former. Aron has narrowed down the 
notion of "elite" to the "minority" that "exercises power" (34, p. 
567) ; elsewhere, he speaks of "ruling classes." Without entering 
into terminological disputes, I propose to examine the general charac- 
teristics ascribed by these three authors to dominating groups and the 
validity of their analyses. 

In their way of posing the problem, the approaches of Pareto, 
Mosca, and Aron entail at many points indications of the sociological 
theory of group conflict as we understand it. All three authors deal 
with the problem of inertia, i.e., the tendency of dominating groups 
to maintain and defend their domination. They also deal with the 
role of legitimacy in the maintenance or change of authority struc- 
tures. Mosca and Pareto, in particular, emphasize the problem of 
social mobility to which we shall have to return. As to the psychology 
of conflict groups, their works contain many a useful suggestion. 
They discuss in some detail the formation and disintegration of 
"aristocracies" as well as other types of social change, basing their 
analyses on thorough historical documentation. If for the discussion 

^® To this list other names would obviously have to be added, among them, above 
all, Max Weber. However, Weber has failed to connect his theory of power and 
authority with the analysis of conflict. Contrary to Aron's, Pareto's, and Mosca 's, his 
woric is suggestive rather than directly indicative of the approach of the present study. 

Social Structure y Group Interests, Conflict Groups 195 

of this section I select only five aspects of the theories of Pareto, 
Mosca, and Aron, it is because this selection is guided by the intention 
to combine a critical examination of these theories with some discus- 
sion of the general characteristics of dominating conflict groups. 

(i) Even in his definition of dominating groups, Aron refers to 
these as "minorities." Mosca does not hesitate to elaborate this into 
the general thesis that the ruling class is "always the less numerous" 
group. The notion of an elite appears to evoke almost automatically 
the idea of the "chosen few," of a small ruling stratum. Thus, even 
Marx describes the action of the proletariat as the "independent move- 
ment of the overwhelming majority in the interest of the overwhelm- 
ing majority" (14, pp. 20 f.), and almost as a matter of course Geiger, 
in his graphical schema of class structure (46, p. 43), represents the 
ruling class by a segment of the whole (circle) much smaller than 
the subjected class. That dominating groups are by comparison with 
their subjected counterpart often insignificantly small groupings is 
an assumption which to my knowledge has never been contested in 
the literature. Not all authors state as clearly as Machiavelli how 
small, exactly, these groups are: "In any city, however it may be 
organized politically, no more than 40 or 50 men attain real power" 
(see 24, p. 271). Mosca, in particular, supplements his political class 
by "another, much more numerous stratum including all those who 
are suited for leading positions" (p. 329) ; but by this extension he 
merely obliterates his analysis without abandoning the minority char- 
acter of elites. In fact, the assumption that in any association the 
number of those subjected to authority is larger than the number of 
those in possession of authority does seem capable of generalization. 
It seems hard to imagine an association in which the "rulers" outweigh 
the "ruled" in number. In every state, the number of cabinet min- 
isters is smaller than the number of citizens j in every enterprise there 
are fewer executives than employees. However, this seemingly gen- 
eral statement requires qualification for industrial societies at an ad- 
vanced stage of development. Today, one is hardly surprised to find 
that in many modern industrial enterprises almost one-third of all 
employees exercise superordinate functions. Delegation of authority 
in industry, in the state, and in other associations makes possible in 
industrial societies dominating groups which are no longer small 
minorities but which in size hardly fall short of subjected groups. 
We have earlier examined some of the problems of delegated au- 
thority and we shall return to this point. By way of generalization, 
these phenomena justify at least the negative statement that it seems 

196 Social Structure y Grouf Interests y Conflict Groups 

to be one of the characteristics of industrial societies that those who 
are plainly subjected to authority in imperatively coordinated associ- 
ations of many types not only do not any longer amount to the "over- 
whelming majority" but actually decrease steadily. Pareto's, Mosca's, 
and Aron's thesis of a small ruling minority requires correction. Le- 
gitimate power may be distributed, if with considerable gradations 
of spheres of authority, over a large number of positions. 

(2) Pareto and Mosca characterize dominating groups by a num- 
ber of peculiar properties which are alleged to be necessary for a group 
to attain and successfully defend its position of power. Pareto em- 
phasizes "energy" and "superiority" (25, p. 230), an "instinct of 
combination," concentration on the proximate, and similar "proper- 
ties" (e.g., pp. 242 f.). Mosca goes even further 5 for him "the ruling 
minorities usually consist of individuals who are superior to the mass 
of the ruled in material, intellectual, and even moral respects, or they 
are at least the descendants of individuals who had such virtues. To 
put it differently, the members of the ruling minority generally have 
real or apparent properties which are highly esteemed and convey 
great influence in their societies" (24, p. SS)- This kind of thesis 
illustrates that pre-sociological character of Mosca's analyses, i.e., the 
speculative recourse from social structures and roles to individuals and 
their "properties," which hardly helps our insight into social relations. 
Without the full consistency of the Aristotelian argument, Mosca 
approximates the notion that certain people are "by nature" rulers or 
ruled, freemen or slaves. This notion, however, in whatever variant 
it may appear, has to be banned radically and finally from the socio- 
logical theory of group conflict. 

Whether dominating conflict groups are characterized by attri- 
butes and patterns of behavior other than common manifest interests 
is a question that can be answered only by empirical observation and 
in relation to specific social conditions. This is in fact the question 
with which we have dealt above in terms of "class culture." It is cer- 
tainly possible that there are societies in which dominating groups are 
also distinguished by patterns of behavior crystallized hypothetically 
in "properties" 5 but it is at least equally possible that the coherence 
of such groups is confined to the defense of common interests within 
well-defined units of social organization without significantly affect- 
ing other spheres of the behavior of the members of ruling groups. 
From the point of view of the theory of group conflict, the "prop- 
erties" of individual group members are in principle indeterminate 
and variable. 

Social Structure y Group Interests, Conflict Groups 197 

(3) Mosca consistently derives from two untenable postulates — 
the minority character of ruling groups and the existence of a common 
culture among them — the conclusion that dominating conflict groups 
are always better organized than subjected groups. "The minority is 
organized simply because it is the minority" (p. SS)- Like its prem- 
ises, this conclusion can by no means be assumed; it is, rather, an 
empirical generalization, and one demonstrably false. Within the 
association of industry, for example, it would appear that there are 
greater obstacles to the formation of an interest group on behalf of 
the incumbents of positions of domination (because of the far-going 
internal differentiation of this quasi-group ? ) than is the case for the 
subjected workers. At the very least, we can say that we know of no 
point of view that would permit the postulate that a transition from 
quasi-groups to interest groups is easier for dominating than for sub- 
jected groups." 

(4) Mosca, and to some extent Pareto, means by the name "ruling 
class" only the incumbents of positions of domination in the political 
society. Pareto recognizes elites in all spheres and associations of 
society, but "governing elites" are for him politically governing elites. 
Mosca limits the field of his analyses by the very concept of "political 
class." It is only Aron who intimates an extension of this approach 
by emphasizing "the distinction between the political power of classes, 
founded on the position occupied in the state by their representatives, 
and their economic power, determined by their place in the process 
of production" (34, p. 572). Yet Aron also presupposes the unity of 
a class ruling in all spheres in which authority is exercised. In so far 
as this presupposition implies a restriction of conflict analysis to the 
association of the political state, it is unnecessary and, indeed, dis- 
advantageous; in so far as it implies the assertion that the "political 
class" is eo ipso the ruling group in all other spheres of society, it is 
once again an untenable empirical generalization. One of the short- 
comings of the theories of Mosca, of Pareto, and, to some extent, of 
Aron is that although these authors derive conflict groups from re- 
lations of authority, they fail to relate these to the crucial category 
of imperatively coordinated associations. 

Ruling groups are, in the first place, no more than ruling groups 

^^ To clarify this problem fully one would have to consider all the conditions of 
organization. Thus it might be feasible to make an empirical generalization to the 
effect that in pre-industrial societies ruling groups were (above all because of easier 
communication) provided with better conditions than subjected groups. In industrial 
societies, however, this clearly does not hold. 

198 Social Structure y Group Interests y Conßict Groups 

within defined associations. In theory, there can be as many com- 
peting, conflicting, or coexisting dominating conflict groups in a society 
as there are associations. Whether and in what way certain associa- 
tions — such as industry and society — are connected in given societies 
is a subject for empirical analysis. Without doubt, such analysis is 
of considerable significance for a theory of conflict. Nevertheless, it 
is analytically necessary and empirically fruitful to retain the possi- 
bility of a competition or even conflict between the ruling groups of 
different associations. In this sense, the expression "ruling class" is, 
in the singular, quite misleading. 

(5) Of the three authors under discussion, Mosca in particular 
has fallen victim to a Marxian overestimation of class analysis. If 
Pareto claims that history is "a cemetery of aristocracies" (25, p. 229), 
he leaves it open whether group conflicts or other forces caused the 
death of ruling elites. But Mosca is quite explicit : "One could explain 
the whole history of civilized mankind in terms of the conflict between 
the attempt of the rulers to monopolize and bequeath political power 
and the attempt of new forces to change the relations of power" (24, 
pp. 64 f.). This is hardly more than a reformulation of the Marxian 
thesis "the history of all hitherto society is the history of class strug- 
gles" (14, p. 6). Mosca's statement is therefore subject to the same 
objections. Ruling groups in the sense of the theory here advanced 
do by no means determine the entire "level of culture of a people" 
(Mosca, 24, p. 54). As coercion theory emphasizes but one aspect of 
social structure, thus the distinction between ruling and subjected 
groups is but one element of society. It would be false to identify 
the upper stratum of a society unequivocally with its ruling conflict 
group. There is no need for these two to be identical with respect to 
their personnel, nor do these categories, even if the personnel of upper 
stratum and ruling conflict group are the same, describe the same 
aspect of social behavior. In any case, ruling classes or conflict groups 
decide not so much the "level of civilization" of a society as the 
dynamics of the associations in which they originated. 

It is a significant if confusing trait of the theories of Pareto and 
Mosca that both of them are concerned less with the explanation of 
social change than with that of stability or, as Pareto explicitly says 
at many points, of "equilibrium." By concentrating their attention 
primarily on the "elite" or "ruling class," they tend to reduce all 

Social Structure y Growp Interests y Conflict Growps 199 

changes to changes in the composition of the ruling class, i.e., to one 
type of social mobility,^^ Pareto's "circulation of elites" and Mosca's 
emphasis on the "ability" of a people "to produce in its womb new- 
forces suited for leadership" (24, p. 227) aim at the same phenom- 
enon, i.e., the regeneration of a leading stratum which is assumed to 
be universally procured by individual mobility. By virtue of this 
emphasis the theories of Pareto and Mosca take a strange turn of 
which their authors are probably not aware. Although both of them 
originally refer to two classes (Pareto, p. 226; Mosca, p. 52), their 
approach gradually and barely noticeably reduces itself to a "one- 
class model," in which only the ruling group functions as a class 
proper. Pareto characteristically speaks, by way of introducing the 
notion of "circulation of elites," of "two groups, the elite and the 
rest of the fofulation" (p. 226), and Mosca similarly distinguishes 
at one point "the subjected masses" and "the political class" (p. 
53).^® Both notions, however — that of a "rest of the population" 
and that of "masses" — are basically residual categories defined by 
privation and not considered as independently operative forces. It 
need hardly be mentioned that this procedure robs any theory of 
conflict of its substance. At this point we see the crucial difference 
between elite theories and conflict theories in the sense of the present 

The almost unnoticed transition from conflict theory to elite 
theory in the works of Pareto and Mosca has one aspect of some sig- 
nificance for our context. This becomes apparent if we contrast this 
modification with Marx^s approach (which at times almost appears 
to commit the opposite mistake and to recognize only the proletariat 
as a class) . The thesis might be advanced that in post-classical history 
of Europe the industrial workers of the nineteenth century consti- 
tuted, indeed, the first subjected group that managed to establish itself 
as such, i.e., that left the stage of quasi-group and organized itself 
as an interest group. Thus, earlier "suppressed classes" could quite 
properly be described as "masses" or "rest of the population," that is, 
as quasi-groups such as the French peasants of Marx's "i8th Bru- 
maire," who provided — as Mosca (p. 104) argues along lines similar 
to Marx's in his study of Louis Bonaparte — merely a basis of legiti- 

^* Quite consistently, then, revolutions are, for Pareto and Mosca, abnormal 
events which betray the weakness of an elite, namely its inability to rejuvenate by 
absorbing new members. 

^® Italics in both quotations mine. 

2CXD Social Structure^ Growp Interests^ Conflict Groups 

macy and "support" of competing "groups within the political class." 
We need not settle this question here. But the fact that it can be 
raised provides a further reason why I have chosen to limit this study 
— contrary to Pareto and Mosca as well as Marx — to industrial so- 
cieties. Perhaps it is feasible to make the general assertion that, in 
principle, ruling and suppressed classes have, in industrial societies, 
equal chances of organization, because in these societies one obstacle 
to the organization of subjected groups characteristic for most earlier 
societies is removed: the impossibility of communication. Although 
I suspect that the theory formulated in this study might be extended 
in such a way as to apply to pre-industrial societies also, I shall confine 
myself to applying it to societies in which manifest conflicts of or- 
ganized interest groups are empirically possible. 

Subjected conflict groups must therefore not be visualized as es- 
sentially unorganized masses without effective force. In analogy to 
the characteristics of ruling groups we can state {a) that they do not 
necessarily comprise the majority of the members of an association, 
{b) that their members are not necessarily connected by "properties" 
or a "culture" beyond the interests that bind them into groups, and 
{c) that their existence is always related to particular associations, so 
that one society may display several subjected conflict groups. Beyond 
these, one distinguishing feature of subjected groups must be empha- 
sized. The Marxian expression "suppressed classes" might appear 
to mean that any such group is characterized by the attributes which 
Marx ascribed to, or found present in, the proletariat of his time. 
However, this implication is by no means intended here. "Pauper- 
ism," "slavery," absolute exclusion from the wealth and liberty of 
society is a possible but unnecessary attribute of the incumbents of 
roles of subjection. Here, again, the connection is indeterminate, i.e., 
variable, and its particular pattern can be established only by empirical 
observation and for particular associations. It is not only conceivable 
that members of the subjected group of one association belong to the 
dominating group of another association, it is above all possible that 
"suppressed classes" enjoy, despite their exclusion from legitimate 
power, an (absolutely) high measure of social rewards without this 
fact impeding their organization as interest groups or their partici- 
pation in group conflicts. Even a "bourgeoisified proletariat" can 
function as a subjected conflict group, for conflict groups and group 
conflicts are solely based on the one criterion of participation in or 
exclusion from the exercise of authority in imperatively coordinated 
associations. Difficult as it may be for minds schooled in Marx to 

Social Structure, Group Interests y Conflict Groups 201 

separate the category of "suppressed class" from the ideas of poverty 
and exploitation, a well-formulated theory of group conflict requires 
the radical separation of these spheres. 


Up to this point I have postponed and at times avoided the ques- 
tion whether the concept of class is a useful concept to employ and, 
if so, what its precise meaning is in the context of the theory of con- 
flict group formation. The reader will not have failed to notice that 
I have in fact strenuously avoided the word "class" in the present 
chapter wherever possible. Before turning now to an attempt to settle 
this rather disturbing question, I want to emphasize one point. In my 
opinion, the problem of the applicability of the concept of class is a 
purely terminological problem. In positive terms, this means that it 
is in part a matter of arbitrary decision, and in part a matter of con- 
venience. Logically, there is no reason why we should not call quasi- 
groups and interest groups classes or anything else. Pragmatically, 
of course, the usage and history of words has to be considered j it is 
unwise to provoke misunderstandings by choosing words which carry 
associations that are not intended. In negative terms, the termino- 
logical nature of this problem means that I see no meaning in the 
statement that class is a "historical concept" in the sense of being in- 
separably tied to a definite historical entity such as the industrial 
proletariat of the nineteenth century. "Historical concepts" of this 
kind are fictions of Hegelianism or, more generally, conceptual real- 
ism. If I shall therefore try to bring together, in the following pages, 
the arguments that can be advanced for and against using the concept 
of class for conflict groups other than those described by Marx, the 
ensuing discussion is concerned exclusively with problems of prag- 
matic convenience, and the conclusion it reaches remains reversible. 

So far in our considerations there have emerged four main reasons 
why the concept of class should not be applied to the analysis of con- 
flicts in post-capitalist societies. The first of these is of a historical 
nature. We have seen that the changes which have occurred since 
Marx's time have in several ways affected the classes with which he 
was concerned. Bourgeoisie and proletariat are no longer uniform 
blocs of identically situated and oriented people if, indeed, they can 
be said to exist at all in post-capitalist society. The progressive insti- 
tutionalization of the values of achievement and equality has removed 
many barriers which for Marx were associated with the concept of 
class. Without anticipating the results of empirical analysis we can 

202 Social Structure y Growp Interests^ Conflict Groufs 

already conclude that conflict groups in modern society are likely to 
be rather loose aggregations combined for special purposes and within 
particular associations. In view of factual developments of this kind, 
it seems certainly questionable whether it is useful to employ for the 
conflict groups of advanced industrial society the concept used for the 
Marxian classes of the nineteenth century. 

This doubt is strengthened by a second argument accruing from 
our theoretical considerations in the present chapter. We have de- 
liberately restricted our model of group formation to elementary and 
highly formal features of the phenomenon. Most of the empirical 
characteristics of conflict groups are subject to a wide range of varia- 
bility the limits of which may be fixed in terms of a constructed model 
but the substance of which needs to be determined by observation and 
experience. Conflict groups may, but need not be, immobile entities j 
they may, but need not be, characterized by a "class culture" j they 
may, but need not, engage in violent conflicts. Moreover, we have 
endeavored to detach the category of conflict groups and the whole 
notion of social conflicts from economic determinants both in the 
Marxian sense of relations of production and ownership and in the 
Weberian sense of socioeconomic class situations. Conceptually, the 
similarity between Marx's and even Weber's concepts of class and 
our concept of conflict group is but slight. There is reasonable doubt 
as to whether there is a chance for the concept of class not to be mis- 
understood if it is applied to conflict groups in the sense of this study. 

Thirdly, in addition to these general conceptual difficulties, the 
question must be raised: what precisely do we mean by class even if 
we decide to apply this term to conflict groups? Are we to follow 
Ginsberg and conceive of classes as quasi-groups, i.e., unorganized 
aggregates of the occupants of positions endowed with role interests? 
Or are we to follow Marx in calling classes only such groups as have 
attained political organization and coherence, and which are interest 
groups? Distinctions such as those between "collectivity" and "class," 
or "class" and "party," or "class in itself" and "class for itself" are 
necessary, but they do not exactly help to render the concept of class 

Finally, the history of the concept in sociological literature has 
to be considered. One may deplore the fact that the terms "class" 
and "stratum" have tended to become interchangeable categories in 
sociological studies, but it remains a fact. While the existence of a 
difference between the study of social conflict and the study of social 
stratification is probably plausible to anybody, the concepts of "class" 

Social Structurey Grouf Interests^ Conßict Groufs 203 

and "stratum," as they are often used today, fail to express this dif- 
ference. Under these conditions, it may not be wise to try to restore 
to the concept of class a meaning which for many it lost long ago. 

There are, on the other hand, three arguments that might be 
held against these doubts about the applicability of the concept of 
class to conflict groups in the sense of our model. First, the alternative 
category of conflict group is so general as to be almost embarrassing. 
We have explicitly distinguished from other conflicts those conflicts 
arising out of the distribution of authority in associations. Yet there 
is no conceivable reason, other than an inconveniently narrow defi- 
nition, why the contestants in conflicts between Protestants and Cath- 
olics, Negroes and whites, town and country should not be called 
conflict groups. Short of using a more specific, but extremely clumsy, 
expression (such as "conflict groups arising from authority structures 
in associations"), the concept of class seems to provide a convenient 
tool for emphasizing the limitations of scope of the theory advanced 
in this study. 

This is, secondly, all the more plausible, since the heuristic pur- 
pose originally associated with the concept of class is also the heuristic 
purpose of this study. When Marx adapted the word "class" to the 
requirements of his theories, he used this word as a term for struc- 
turally generated groups that engage in conflicts over existing arrange- 
ments of social structure. It is true that before Marx the term "class" 
was used by a number of authors in a rather less specific sense j but 
it is probably fair to say that it was Marx's category which became 
germinal for later students in the field and which therefore represents 
its original version. The essential importance of this heuristic purpose 
has been emphasized at many points in our considerations. Since there 
is no other concept that expresses this purpose with equal clarity, one 
might consider it reasonable to retain the concept of class despite all 
qualifications necessitated by the arguments against it. 

One of these arguments has referred to the history of the concept 
in sociological literature. Thirdly, however, there is one not entirely 
insignificant branch of sociological thinking which has consistently 
used (and uses) the term "class" in the form, if not the substance, 
assigned to it by Marx. This is true not only for many Marxist 
scholars whose work is, as we have seen, often pitifully barren and 
fruitless, but also for eminent non-Marxist (although possibly Marx- 
ian) sociologists such as Renner and Geiger, Aron and Gurvitch, 
Pareto and Mosca, Marshall and Ginsberg, Lipset and Bendix, and 
nany others. We might go even further and assert that the trend 

204 Social Structure y Group InterestSy Conflict Groups 

of conceptual development in the work of these scholars anticipates 
in many ways the theses advanced in the present study. Many of 
them have tried to refine the concept of class by maintaining its heu- 
ristic purpose while altering its substance} quite often, this altering 
of substance meant a shift from property to power as a determinant, 
or other attempts at generalization. In using the concept of class for 
Marx's bourgeoisie and proletariat as well as for modern and utterly 
diflFerent conflict groups, one could refer not only to the origin of 
this concept with Marx, but also to a great and unbroken tradition in 
sociological analysis. 

It is hard to weigh the "pros and cons" of the preceding argument 
entirely rationally j an element of personal preference will probably 
enter into any decision. Without trying to argue for this decision at 
any length, I will therefore state immediately that in my opinion the 
case in favor of retaining the concept of class is still sufficiently strong 
to warrant its application to even the most advanced industrial so- 
cieties. This decision does involve, of course, a polemical stand against 
all those who "falsify" the term "class" by applying it to what should 
properly be called social strata. It also involves considerable exten- 
sions of the concept as it was used by Marx as well as by all Marxists 
and Marxians. But it emphasizes that in class analysis we are con- 
cerned {a) with systematic social conflicts and their structural origin, 
and {b) with but one specific type of such conflicts. 

In terms of our model, the term "class" signifies conflict groups 
that are generated by the differential distribution of authority in im- 
peratively coordinated associations. This definition implies no as- 
sumption as to the looseness or rigidity of their coherence, the pres- 
ence or absence of a common culture or ideology (beyond specific 
interests) among their members, and the intensity or lack of intensity 
of their engagement in social conflicts. 

It will be noted that this definition is inconclusive with respect 
to the differentiation of quasi-groups and interest groups. I would 
suggest that it is useful to leave it so. The category of class is a general 
term for groupings of the kind described more specifically in our 
model of conflict group formation. For all particular purposes of 
analysis, it is necessary to abandon this general category in favor of 
the more specific concepts of quasi-group and interest group. The 
attempt to confine the concept of class to either of these is bound, 
indeed, to provoke misunderstandings. Classes, like conflict groups, 
indicate an area and type of sociological analysis rather than its sub- 

Social Structure, Group Interests y Conflict Groufs 205 

stance. Both terms are more useful in compounds such as "class 
analysis," "class structure," or "class conflict" than on their own. This 
is but one further illustration of the essential insignificance of a ter- 
minological dispute about these matters. For purposes of the present 
study, and without any dogmatic insistence on terms, I propose to 
dissolve the alternative "classes or conflict groups" into the definition 
"classes as conflict groups." 


Conflict Groups, Group Conflicts, and 
Social Change 


Classes, understood as conflict groups arising out of the authority 
structure of imperatively coordinated associations, are in conflict. 
What are — so we must ask if we want to understand the lawfulness 
of this phenomenon — the social consequences, intended or unintended, 
of such conflicts? The discussion of this question involves, almost 
inevitably, certain value judgments. I think that R. Dubin is right 
in summarizing at least one prominent attitude toward the functions 
of social conflict as follows: "From the standpoint of the social order, 
conflict is viewed from two positions: {a) it may be destructive of 
social stability and therefore 'bad' because stability is goodj {b) it 
may be evidence of the breakdown of social control and therefore 
symptomatic of an underlying instability in the social order. Both 
positions express a value preference for social stability" (77, p. 183). 
I would also agree with Dubin's own position: "Conflict may be 
labeled dysfunctional or symptomatic of an improperly integrated 
society. The empirical existence of conflict, however, is not challenged 
by the stability argument. . . . The fact of the matter is that group 
conflict cannot be wished out of existence. It is a reality with which 
social theorists must deal in constructing their general models of social 
behaviour" (p. 184). But I think that in two respects Dubin might 
have been rather less cautious. First, I should not hesitate, on the 
level of value judgments, to express a strong preference for the con- 
cept of societies that recognizes conflict as an essential feature of their 
structure and process. Secondly, and quite apart from value judg- 
ments, a strong case can be made for group conflict having conse- 
quences which, if not "functional," are utterly necessary for the social 
process. This case rests on the distinction between the two faces of 
society — a distinction which underlies our discussions throughout this 
study. It is perhaps the ultimate proof of the necessity of distinguish- 

Confiia Groups, Group Conflicts, and Social Change 207 

ing these two faces that conflict itself, the crucial category in terms 
of the coercion model, has two faces, i.e., that of contributing to the 
integration of social "systems" and that of making for change. 

Both these consequences have been admirably expressed by L. 
Coser. (Although, to my mind, Coser is rather too preoccupied with 
what he himself tends to call the "positive" or "integrative func- 
tions" of conflict.) On the one hand, Coser states in the unmistakable 
terminology of the integration theory of society (for which see my 
italics): "Conflict may serve to remove dissociating elements in a 
relationship and to re-establish unity. Insofar as conflict is the reso- 
lution of tension between antagonists it has stabilizing functions and 
becomes an integrating component of the relationship. However, not 
all conflicts are positively functional for the relationship. . . . Loose- 
ly structured groups, and open societies, by allowing conflicts, institute 
safeguards against the type of conflict which would endanger basic 
consensus and thereby minimize the danger of divergences touching 
core values. The interdependence of antagonistic groups and the 
crisscrossing within such societies of conflicts, which serve to ^sew the 
social system, together"* by cancelling each other out, thus prevent dis- 
integration along one primary line of cleavage" (81, p. 80). On the 
other hand, Coser follows Sorel in postulating "the idea that conflict 
. . . prevents the ossification of the social system by exerting pres- 
sure for innovation and creativity" and states: "This conception seems 
to be more generally applicable than to class struggle alone. Conflict 
within and between groups in a society can prevent accommodations 
and habitual relations from progressively impoverishing creativity. 
The clash of values and interests, the tension between what is and 
what some groups feel ought to be, the conflict between vested interests 
and new strata and groups demanding their share of power, wealth 
and status, have been productive of vitality" (80, pp. 197 f.). 

Conflict may, indeed, from a Utopian point of view, be conceived 
as one of the patterns contributing to the maintenance of the status 
quo. To be sure, this holds only for regulated conflicts, some of the 
conditions of which we shall try to explore presently. Coser's analysis 
of Simmel (81) has convincingly demonstrated that there is no need 
to abandon the integration theory of society simply because the phe- 
nomenon of conflict "cannot be wished away" but is a fact of obser- 
vation. In this sense, conflict joins role allocation, socialization, and 
mobility as one of the "tolerable" processes which foster rather than 
endanger the stability of social systems. There seems little doubt, 

208 Conflict GrowpSy Group Conflicts j and Social Change 

however, that from this point of view we can barely begin to under- 
stand the phenomenon of group conflicts. Were it only for its "posi- 
tive functions," for which Coser found so many telling synonyms, 
class conflict would continue to be rather a nuisance which the sociolo- 
gist would prefer to dispense with since it may, after all, "endanger 
basic consensus." So far as the present study is concerned, "continu- 
ing group conflict" will be regarded as "an important way of giving 
direction to social change" (Dubin, 77, p. 194). Societies are essen- 
tially historical creatures, and, because they are, they require the 
motive force of conflict — or, conversely, because there is conflict, there 
is historical change and development. The dialectics of conflict and 
history provide the ultimate reason of our interest in this phenomenon 
and at the same time signify the consequences of social conflict with 
which we are concerned. 

Dubin's observation that conflict is a stubborn fact of social life 
is undoubtedly justified. Earlier, we have made the assertion explicit 
that social conflict is ubiquitous 5 in fact, this is one of the premises 
of our analysis. Possibly, this premise permits even further generali- 
zation. There has been in recent years some amount of interdiscipli- 
nary research on problems of conflict. In specific features the results 
of these interdisciplinary eflForts remain as yet tentative j but one con- 
clusion has been brought out by them with impressive clarity: it 
appears that not only in social life, but wherever there is life, there 
is conflict.^ May we perhaps go so far as to say that conflict is a con- 
dition necessary for life to be possible at all? I would suggest, in any 
case, that all that is creativity, innovation, and development in the 
life of the individual, his group, and his society is due, to no small 
extent, to the operation of conflicts between group and group, indi- 
vidual and individual, emotion and emotion within one individual. 
This fundamental fact alone seems to me to justify the value judg- 
ment that conflict is essentially "good" and "desirable." 

If I here assume social conflict, and the particular type of group 
conflict with which we are concerned in the present study, to be ubiq- 

■^ This and numerous other statements in the present chapter are based on dis- 
cussions with and publications of psychologists, anthropologists, lawyers, and social 
psychologists at the Center for Advanced Study in the Behavioral Sciences, Stanford, 
California. John Bowlby, M.D., and Professor Frank Newman, LL.D., have been 
particularly helpful in making suggestions. In support of the statement in the text I 
might also refer, however, to the symposium published in Conflict Resolution (77), 
which includes contributions by economists, sociologists, social psychologists, anthro- 
pologists, and psychologists, and strongly supports my point. 

Conflict Groups, Group Conflicts, and Social Change 209 

uitous, I want this statement to be understood more rigidly than is 
usual. At an earlier point I have intimated what I mean by rigidity 
in this sense. One or two remarks in addition to these earlier hints 
seem in order. In summarizing earlier research, Mack and Snyder 
state with some justice that by most authors "competition is not re- 
garded as conflict or a form of conflict" (77, p. 217). The alleged 
difference between the two is identified differently by different au- 
thors. T. H. Marshall emphasizes common interests, rather than 
divergent interests, as characteristic of states of competition or con- 
flict (59, p. 99). For Mack and Snyder, "competition involves striv- 
ing for scarce obj ects . . . according to established rules which strict- 
ly limit what the competitors can do to each other in the course of 
striving 5 the chief objective is the scarce object, not the injury or 
destruction of an opponent per se" (77, p. 217). It seems to me, 
however, that it is not accidental if Mack and Snyder state a little later 
that "conflict arises from 'position scarcity' and 'resource scarcity,' " 
and that therefore "conflict relations always involve attempts to gain 
control of scarce resources and positions" (pp. 218 f.). Despite ter- 
minological traditions, I can see no reason why a conceptual distinc- 
tion between competition and conflict should be necessary or, indeed, 
desirable.^ Like competition, conflict involves a striving for scarce 
resources. From the point of view of linguistic usage, it is perfectly 
proper to say that conflicting interest groups compete for power. As 
far as the "established rules" of competition are concerned, they 
emphasize but one type of conflict, namely, regulated conflict. In the 
present study, the notion of conflict is intended to include relations 
such as have been described by many other authors as competitive. 
Another distinction almost general in the literature is that between 
changes "within" and changes "of" or conflicts "within" and conflicts 
"about" the system. Many authors have been at pains to define these 
differences. Coser, e.g., proposes "to talk of a change of system when 
all major structural relations, its basic institutions and its prevailing 
value system have been drastically altered," but admits that "in con- 
crete historical reality, no clear-cut distinctions exist" (80, p. 202). 
Marshall distinguishes more specifically "conflict that arises out of 

^ At least, no such reason has been put forward. It might be argued, of course, 
that the concept of competition employed in economic theory is rather different from 
that defined by Marshall or Mack and Snyder, and does not carry any conflict connota- 
tion. I am not entirely sure that this argument is justified, but for purposes of the 
present analysis competition in a technical economic sense will be excluded. 

210 Conflict Growps, Growp Conflicts ^ and Social Change 

the division of labor, conflict, that is to say, over the terms on which 
cooperation is to take place, as illustrated by a wage dispute between 
employer and employed," from "conflict over the system itself upon 
which the allocation of functions and the distribution of benefits are 
based" (59, p. 99). Thinking in terms of inclusive epochs like "feu- 
dalism" and "capitalism" as well as in terms of the existence of po- 
litical parties that propose to change "the whole system" can probably 
explain the widespread feeling that a distinction between "changes 
within" and "changes of" is necessary. But apart from these, it is 
surely no coincidence that it was Parsons who emphasized that "it is 
necessary to distinguish clearly between the processes within the sys- 
tem and processes of change of the system." This very distinction 
betrays traces of the integration approach to social analysis. If con- 
flict and change are assumed to be ubiquitous, there is no relevant 
difference between "changes within" and "changes of," because the 
"system" is no longer the frame of reference. It may be useful to 
distinguish more or less intense or violent conflicts and major and 
minor changes, but these are gradations to be accounted for in terms 
of intervening variables of an empirical nature. In the present study, 
no assumption is implied as to the type of change or conflict effected 
by the antagonism of conflict groups. Wage disputes as well as po- 
litical conflicts "over the system itself" will be regarded as mani- 
festations of class conflict, i.e., of clashes of interest arising out of 
and concerned with the distribution of authority in associations. 

As with the theory of class formation, the real problems of the 
theory of class conflict consist in the identification of the empirical 
variables delimiting the range of variability of forms and types. 
Change and conflict are equally universal in society. But in historical 
reality we always encounter particular changes and specific conflicts, 
and these, even in the more limited sphere of class conflict, present 
a varied picture of manifold types and forms. Assuming the ubiquity 
of conflict and change, we have to try to discover some of the factors 
that influence its concrete shapes. 


The substance of the theory of class action, or class conflict, can 
be summarized in one statement: conflict groups in the sense of this 
study, once they have organized themselves, engage in conflicts that 
effect structure changes. The theory of class action presupposes the 
complete formation of conflict groups and specifies their interrelations. 

Conflict Groups J Group Conflicts y and Social Change 211 

However, this tautological statement is evidently not all that can be 
said about group conflicts, nor is it all that one would expect a theory 
of group conflict to provide. Beyond a basic assumption of this kind, 
a theory of class conflict has to identify and systematically interrelate 
those variables that can be shown to influence patterns of intergroup 
conflict. In the present chapter several such variables will be discussed 
in some detail, their selection being guided by the significance they 
suggest for the course and outcome of class conflict. Before we embark 
upon this discussion, however, there is one preliminary question that 
has to be settled. The statement that class conflicts are empirically 
variable is sufficiently vague to be almost meaningless. What is it — 
we must ask — about class conflicts that is variable and therefore sub- 
ject to the influence of factors to be identified? In this question, the 
categories of intensity and violence are essential. In some connection 
or other, the terms "intensity" and "violence" can be found present 
in any discussion of conflict. Here is one example. Mack and Snyder, 
in their summary of earlier research, on the one hand derive the 
proposition "a high degree of intimacy between the parties, as con- 
trasted with a high degree of functional interdependence, will in- 
tensify conflict" (77, p. 225), while, on the other hand, they suggest 
"the more integrated into the society are the parties to conflict, the 
less likely will conflict be violent" (p. 227). The distinction between 
the two concepts is not perhaps entirely clear from these statements, 
and, indeed, many authors use them almost synonymously. Yet there 
is an important difference between them, as Simmel knew when he 
said: "It is almost inevitable that an element of commonness injects 
itself into . . . enmity once the stage of open violence yields to an- 
other relationship, even though this new relation may contain a com- 
pletely undiminished sum of animosity between the two parties" (see 
81, p. 121). That conflict is variable means that its intensity and vio- 
lence are variable j but the two may vary independently and are, there- 
fore, distinct aspects of any conflict situation.^ 

The category of intensity refers to the energy expenditure and 
degree of involvement of conflicting parties. A particular conflict may 
be said to be of high intensity if the cost of victory or defeat is high 
for the parties concerned. The more importance the individual par- 
ticipants of a conflict attach to its issues and substance, the more intense 
is this conflict. For class conflict a continuum might be constructed 
ranging, e.g., from a conflict within a chess club which involves but 

* All italics in the quotations of this paragraph are mine. 

212 Conflict Groups, Group Conflicts, and Social Change 

a small segment of the individual personalities concerned to the over- 
riding class conflict, in Marx's analyses, in which individuals are en- 
gaged with almost their entire personalities. In operational terms, 
the cost aspect is here crucial. Members of a group that strives to 
upset the authority structure of a chess club stand to lose less in case 
of defeat than members of a trade union who endeavor to change the 
authority structure of the enterprise (or their own social conditions 
by way of this authority structure) .* The cost of defeat, and with it 
the intensity of conflict, differs in these cases. 

By contrast to its intensity, the violence of conflict relates rather 
to its manifestations than to its causes j it is a matter of the weapons 
that are chosen by conflict groups to express their hostilities. Again, 
a continuum can be constructed ranging from peaceful discussions to 
militant struggles such as strikes and civil wars. Whether or not class 
conflict expresses itself in militant clashes of interest is in principle 
independent of the intensity of involvement of the parties. The scale 
of degree of violence, including discussion and debate, contest and 
competition, struggle and war, displays its own patterns and regulari- 
ties.^ Violent class struggles, or class wars, are but one point on this 

While violence and intensity of conflict vary independently, sev- 
eral of the factors shortly to be discussed affect both. This fact can 
be illustrated with reference to one factor which has been mentioned 
already and which need not therefore be discussed again at any length. 
I have mentioned in the preceding chapter that the conditions of 
organization of interest groups continue to affect group conflict even 
after the complete formation of conflict groups. They are, in this 
sense, a factor which, among others, accounts for variations of inten- 
sity and violence. With respect to the intensity of class conflict, the 
political conditions of organization appear especially relevant. It may 
be suggested that, for the individuals concerned, involvement in con- 

* I have as yet not given a systematic exposition of the patterns of change effected 
by class conflict; the formulation in the text may therefore give rise to misunderstand- 
ings. These will, I hope, be cleared up in the section on "class conflict and structure 
change" later in this chapter. 

° In terms of the distinction thus introduced, we are now able to reformulate the 
contrast between the conception of conflict here assumed and that of several other 
authors. The latter tend to confine the term "conflict" to one point on the scale of 
degree of violence, namely, highly violent clashes. In the present study, however, con- 
flict is conceived as including the whole scale, i.e., any clash of interest independent 
of the violence of its expressions. 

Conflict Groups y Grou-p Conflicts ^ and Social Change 213 

flicts decreases as the legitimacy of conflicts and, by implication, their 
issues become recognized. However, in the ensemble of factors af- 
fecting intensity of conflict, the specific weight of the conditions of 
organization is probably not very great. By contrast, it is considerable 
among the variables involved in determining the violence of conflict 
manifestations. As soon as conflict groups have been permitted and 
been able to organize themselves, the most uncontrollably violent 
form of conflict, that of guerrilla warfare, is excluded. Moreover, the 
very fact of organization presupposes some degree of recognition 
which in turn makes the most violent forms of conflict unnecessary 
and, therefore, unlikely. This is not to say, of course, that conflicts 
between organized groups cannot be highly intense and violent. The 
conditions of organization are but one, and not the most important, 
factor among many. Of these I have selected four which seem to me 
of particular importance and which will be dealt with separately in 
the following sections of this chapter. 


One of the crucial elements of the theory of group conflict con- 
sists in the strict relation of conflicts to particular associations. Any 
given conflict can be explained only in terms of the association in which 
it arose and, conversely, any given association can be analyzed in terms 
of the conflicts to which it gives rise. In theory, this approach would 
suggest that inclusive societies present the picture of a multitude of 
competing conflicts and conflict groups. The two-class model applies 
not to total societies but only to specific associations within societies 
(including, of course, the inclusive association of the state, i.e., the 
whole society in its political aspect). If, in a given society, there are 
fifty associations, we should expect to find a hundred classes, or con- 
flict groups in the sense of the present study. Apart from these, there 
may be an undetermined number of conflict groups and conflicts 
arising from antagonisms other than those based on the authority 
structure of associations. In fact, of course, this extreme scattering 
of conflicts and conflict groups is rarely the case. Empirical evidence 
shows that different conflicts may be, and often are, superimposed in 
given historical societies, so that the multitude of possible conflict 
fronts is reduced to a few dominant conflicts. I suggest that this phe- 
nomenon has considerable bearing on the degree of intensity and vio- 
lence of empirical conflicts. 

214 Conflict Groups, Group Conflicts y and Social Change 

The pluralism-superimpositlon scale which might thus be con- 
structed has two distinct dimensions. One of these relates to the sepa- 
ration or combination of conflicts of the class type in different asso- 
ciations. Let us restrict ourselves, for purposes of illustration, to the 
three associations of the state, industry, and the church in countries 
in which one church dominates the sphere of religious institutions. 
It is conceivable that the ruling and the subjected groups of each of 
these associations are largely separate aggregations. The dignitaries 
of the church may be mere citizens of the state and may have no in- 
dustrial property or authority. Similarly, the citizens of the state may 
be church dignitaries or industrial managers. This is the kind of situ- 
ation here described as pluralistic. Within each of the three associa- 
tions there are (class) conflicts, but, as between these, there is dissoci- 
ation rather than congruence. Evidently, complete dissociation and 
pluralism are, in the case mentioned, empirically rather unlikely. It 
is more probable that the workers of industry are at the same time 
mere members of the church and mere citizens of the state. One might 
expect that the dignitaries of the church are in some ways connected 
with the rulers of the state and possibly even with the owners or man- 
agers of industry. If this is the case, (class) conflicts of different 
associations appear superimposed j i.e., the opponents of one associa- 
tion meet again — with different titles, perhaps, but in identical re- 
lations — in another association. In this case, the personnel of the 
conflict groups of different associations is the same. 

Such congruence may also occur with conflict groups of different 
types. Again, a realistic example may serve to illustrate the point. 
We might suppose that in a given country there are three dominant 
types of social conflict: conflict of the class type, conflict between town 
and country, and conflict between Protestants and Catholics. It is of 
course conceivable that these lines of conflict cut across each other in 
a random fashion, so that, e.g., there are as many Protestants among 
the ruling groups of the state as there are Catholics and as many towns- 
people in either denomination as there are countrypeople. However, 
here, too, we might suspect that dissociation and pluralism are em- 
pirically rather unlikely to occur. One would not be surprised to find 
that most Protestants live in towns and most Catholics in the country, 
or that only one of the denominations commands the instruments of 
political control. If this is so, we are again faced with a phenomenon 
of superimposition in the sense of the same people meeting in different 
contexts but in identical relations of conflict. 

With respect to the violence of manifestations of conflict, the 
pluralism-superimposition scale is not likely to be a factor of great 

Conflict Groups, Group Conflicts, and Social Change 215 

significance. While there is a possible (negative) correlation between 
the degree of pluralism and the violence of conflicts in a given society, 
there is little reason to believe that dissociation of types and contexts 
of conflict makes industrial strikes, for example, impossible. Only in 
the inclusive association of the state would there seem to be a proba- 
bility of pluralism reducing and superimposition increasing the vio- 
lence of interest clashes. 

At the same time, this scale is of the utmost importance for vari- 
ations in the intensity of class conflict. The proposition seems plausi- 
ble that there is a close positive correlation between the degree of 
superimposition of conflicts and their intensity. When conflict groups 
encounter each other in several associations and in several clashes, the 
energies expended in all of them will be combined and one over- 
riding conflict of interests will emerge. The situation with which 
Marx dealt is a case in point. If incumbents of subjected positions 
in industry are also subjected in all other associations j if they are, 
moreover, identical with conflict groups other than those determined 
by authority relations, a "division of society into two large hostile 
classes" may indeed result — a situation, that is, in which one inclusive 
conflict dominates the picture of the total society. If, on the other 
hand, the inevitable pluralism of associations is accompanied by a 
pluralism of fronts of conflict, none of these is likely to develop the 
intensity of class conflicts of the Marxian type. There is in this case, 
for every member of the subjected class of one association, the promise 
of gratification in another association. Every particular conflict re- 
mains confined to the individual in one of his many roles and absorbs 
only that part of the individuaPs personality that went into this role.® 
The empirical analysis of pluralism and superimposition of contexts 
and types of conflict is one of the important problems suggested by 
the theory of social classes and class conflicts. 


In connection with the concept of class situation, we have briefly 
(and, for the most part, critically) considered the relation between 
class structure and social stratification at several points in the pre- 

® This type of analysis seems to me to provide one of the answers to the question 
why there is no socialism in the United States. Throughout her history, the pluralism 
of associations and conflicts has made inclusive conflict groups held together by quasi- 
religious ideologies unnecessary. There has been no single group that enjoyed uni- 
versal privilege or sufi"ered universal alienation. 

2 1 6 Conßict GroufSy Group ConßictSy and Social Change 

ceding chapters. It is not my intention to repeat here what has been 
said before. Rather, I propose to summarize and extend these earlier 
discussions with particular emphasis on the problems of intensity and 
violence of class conflicts. It is evident that in the context of a theory 
of group conflict of the type under discussion, "class situation" is an 
unnecessary concept. It means no more than what we have described 
as the authority position of aggregates in associations. The condition 
of a quasi-group in terms of the distribution of authority signifies the 
"situation" that underlies class conflict. However, the traditional con- 
cept of class situation includes a number of elements which, while 
irrelevant for the formation of social classes, affect their patterns of 
conflict in ways to be defined. Property, economic status, and social 
status are no determinants of class, but they do belong to the factors 
influencing the empirical course of clashes of interest between conflict 

As with contexts and types of conflict, the problem of rewards and 
facilities can be seen in terms of a contrast between divergence and 
parallelism, or pluralism and superimposition. Thus, property can, 
but need not, be associated with the exercise of authority. It is con- 
ceivable that those who occupy positions of domination in industry 
do not own industrial property — and, indeed, that those in positions 
of subjection do own such property. The separation of ownership 
and control, and certain systems of the distribution of shares to in- 
dustrial workers, are cases in point. While neither of these structural 
arrangements eliminates the causes of (industrial) conflict, they have 
an impact on its intensity and violence. Once again, a certain paral- 
lelism between authority and property ownership may seem more 
probable, but it is not necessary. 

The same holds for the economic status of persons in different 
authority positions. By economic status I shall here understand status 
in terms of strictly occupational rewards such as income, job security, 
and general social security as it accrues from occupational position. It 
is both possible and reasonably probable that those in positions of 
domination enjoy a somewhat higher economic status, and that these 
two attributes of social position are in this sense superimposed. But 
numerous illustrations could also be given for divergences between 
the two. In the early labor unions, and for many shop stewards and 
local union secretaries today, authority involves a comparative loss of 
income and security. In the Roman Catholic church, authority is sup- 
posed, in theory if not in practice, to be accompanied by low economic 
status. In totalitarian countries, political authority usually conveys 

Conflict Groups, Group ConfliciSy and Social Change 217 

high incomes but also a high degree of insecurity which lowers the 
economic status of dominant groups. Such divergences of authority 
position and economic status make for a plurality of noncongruent 
scales of position in a society, which constitutes one of the critical 
facts of class analysis. 

Divergences of position are even more evident if we contrast 
authority positions with people's social status in the sense of the 
prestige attached to their position by themselves and by others in 
relevant universes of ranking. The prestige of power is a highly pre- 
carious quantity in all societies. Unless all existing studies are wrong 
in their findings, there would in fact seem to be, for persons in the 
upper ranges of the status scale, an inverse relation between the au- 
thority and the prestige. The judge (United States), the doctor 
(Britain), and the university professor (Germany) enjoy a markedly 
higher prestige than the cabinet minister or the large-scale entrepre- 
neur.^ Probably, the theory of class conflict with its assumption of 
opposing role interests would account for this phenomenon. On the 
other hand, there are and have been associations in which the division 
of authority and the scale of prestige followed identical lines. In the 
industrial enterprise, this would still seem to be the case in most 
countries (and with the possible exception of scientifically trained 
staff members). Thus, we also find here an empirically variable re- 
lation that is likely to affect the course of class conflict. 

All examples chosen in the preceding paragraphs serve to illus- 
trate the phenomenon of relative deprivation, i.e., the situation in 
which those subjected to authority are at the same time relatively 
worse placed in terms of socioeconomic status. However, in nine- 
teenth-century Europe, and in some countries even today, we en- 
counter what by contrast may be called an absolute deprivation of 
groups of people in socioeconomic terms. If the social condition of 
industrial workers, who are as such excluded from authority, falls 
below a physiological subsistence minimum or "poverty line," the 
effects of such deprivation are likely to be different in kind from those 
of relative deprivation. I would suggest that in this case, and in this 
case only, the superimposition of scales of status and the distribution 
of authority is likely to increase the violence of class conflict. This 
is a subtle and complex relation. So far as we know, oppression and 
deprivation may reach a point at which militant conflict motivation 
gives way to apathy and lethargy. Short of this point, however, there 

^ For relevant data, cf. the studies by the National Opinion Research Center 
(120), Glass (107), and Bolte (103). 

2 1 8 Conflict Groups, Group Conflicts , and Social Change 

is reason to believe that absolute deprivation coupled with exclusion 
from authority makes for greater violence in conflict relations. 

Relative deprivation, on the other hand, tends to affect the in- 
tensity of conflict rather than its violence. If incumbents of positions 
of subjection enjoy the countervailing gratification of a relatively high 
socioeconomic status, they are unlikely to invest as much energy In 
class conflicts arising out of the authority structure of associations as 
they would If they were deprived of both authority and socioeconomic 
status. Dominant groups are correspondingly not so likely to be as 
involved in the defense of their authority unless their high socio- 
economic status Is simultaneously Involved. In terms of the Intensity 
of conflict, pluralism would again seem to make for a decrease, and 
superimposition or congruence for an increase:^ the lower the corre- 
lation is between authority position and other aspects of socioeconomic 
status, the less Intense are class conflicts likely to be, and vice versa. 


Since Marx, the Idea of a "classless society" has remained an often- 
used category In sociological literature. By "Marxist" scholars it Is 
applied to a number of existing societies as an allegedly valid category 
of description. But among "non-Marxist" social scientists, too, the 
concept of classless society is occasionally used for describing empirical 
states of society. Thus S. Landshut (53) has tried to demonstrate 
the classlessness of present-day Western industrial societies. In a 
rather more definite sense, a number of sociologists have employed 
the concept of a classless society to describe more limited phenomena, 
such as the agricultural cooperatives of Israel. J. Ben-David, for 
example, speaks on this basis of a "collectlvlst," "classless" stage of 
the social development of Israel (125). We have now assembled 
the materials for examining the sense and nonsense of the sociological 
category of a classless society. This examination will reveal an addi- 
tional factor affecting the intensity and violence of class conflict. 

On the basis of the assumptions and models introduced in this 

^ This proposition must be opposed to the assumption of integration theorists that 
the congruence of different scales of social position is a requisite of stable, integrated 
societies (cf. Parsons, in 35). The exact opposite seems true, even from the point 
of view of integration theory. I cannot help feeling that this is one of the points at 
which integration theorists display — unwillingly, to be sure — almost totalitarian con- 

Conflict Groups y Grouf Conflicis, and Social Change 219 

study, the concept of classless society can mean either of two things. 
First, it may be intended to describe societies in which there are no 
structures of authority that give rise to the formation of conflict groups 
and group conflicts. A society is classless if it is "powerless," i.e., if 
there is no authority exercised in it at all, or if such authority is dis- 
tributed equally among all citizens. But in this sense the category 
of classless society is sociologically meaningless. It may be possible 
to conceive of a society in which all differences of income and f restige 
are leveled and which is therefore "stratumless," but it is hardly 
possible to imagine a society in which there is no differentiation of 
roles in terms of legitimate 'power. Permanent anarchism is socially 
Utopian. Any society, and, indeed, any social organization, requires 
some differentiation into positions of domination and positions of sub- 
jection. No matter what the formal nature of the authority mecha- 
nism, it is a functional imperative of social organizations. Since classes 
can be explained in terms of the differential distribution of authority, 
there is no sociological substance in the assumption of a classless society 
devoid of differentiated authority structures. 

However, the idea of a classless society may be understood in a 
second sense. It is possible to conceive of a society whose structure 
contains positions equipped with different authority rights but which 
does not enable any group of persons to occupy these positions regu- 
larly and exclusively. The same might hold in imperatively coordi- 
nated associations other than the state, e.g., in industry. Associa- 
tions may be governed by the principle of an alternating chair- 
manship, according to which the incumbency of positions of domi- 
nation may or may not be patterned. The collective settlements 
{kibbutzim) of Israel seem to provide a case in point. At least origi- 
nally it was stipulated that every member in turn was to occupy the 
positions of leadership for relatively short periods of time.® In view 
of examples of this kind, it seems plausible to argue that where there 
is no group which is capable of monopolizing the positions of authority, 
it is virtually impossible for coherent conflict groups to emerge, and 
the society or association in question is therefore classless. To be sure, 
this is a kind of classlessness rather different from that of the Utopian 
anarchy i still, it cannot be denied that it makes sense to speak of class- 

® The same principle (of "annuity") may also be found in other organizations, 
such as in German universities where the administrative and scholastic head {Rektor) 
changes every year, and every full professor must (or may) expect to be elected in his 

220 Conflict Groups y Group Conflicts, and Social Change 

lessness in this case also. We might say that societies and associations 
governed by a permanently "alternating chairmanship" are classless 
so far as social mobility is concerned, for it is not the structure of 
positions but the fluctuation of personnel that in this case prevents 
the formation of classes and conflict between them. The example of 
classlessness by fluctuation provides a welcome opportunity to try to 
settle the intricate problem of the relationship of social mobility to 
class conflict. 

Here, as elsewhere, the concept of social mobility is too general 
to be useful. Different types of mobility have to be distinguished, 
and their relation to class conflict examined separately. For purposes 
of this analysis, it seems sufficient to distinguish between intergenera- 
tion mobility — i.e., fluctuations that from the individual's point of 
view occur at the beginning of his occupational (or even educational) 
career — and intrageneration mobility, i.e., fluctuations during the oc- 
cupational life of the individual. Either of these, if present to any 
considerable extent, of course characterizes societies in which class 
membership is not an inescapable and inherited fate. We shall pres- 
ently have to return to this other extreme of the mobility scale which, 
according to our theory, suggests a very high intensity of class conflict. 

Intergeneration mobility seems fully compatible with class for- 
mation and class conflict. If a man's position in the authority structure 
of an association remains the same throughout his membership in this 
association, it appears likely that he belongs to a quasi-group as well 
as to an interest group growing out of this, even if his son or his father 
belongs to a diff^erent class. Schumpeter's comparison of classes with 
"a hotel or an autobus" which "are always occupied, but always by 
different people," is here pertinent (27, p. 171). Where the person- 
nel of classes changes between generations only, there is a sufficient 
degree of stability to permit the formation of conflicting interest 
groups. Janowitz's finding that intergeneration mobility has no detri- 
mental consequences for the coherence of secondary groupings may 
be regarded as an empirical confirmation of this thesis. 

The case of societies in which there is a high degree of intra- 
generation mobility is rather more difiicult. To begin with, further 
distinctions are here required. Not all types of intrageneration mo- 
bility afi^ect class formation and class conflict. At least potentially, 
classes are large groupings which may display, from the point of view 
of social stratification, considerable differentiation within themselves. 
Mobility within classes, however, is entirely irrelevant for our con- 
text. Thus, upward and downward movements between skilled, semi- 

Conflict Groups, Group Conflicts, and Social Change ii\ 

skilled, and unskilled industrial occupations do not affect the stability 
of the conflict group of industrial workers. Moreover, single upward 
or downward moves by individuals, even if they involve a change of 
class allegiance, do not appear destructive of classes. It is only the 
institutionalized principle of an alternating chairmanship that may 
give rise to a state of quasi-classlessness. If the individual can change 
his class belongingness at will, or is even forced to do so regularly j 
if, e.g., the worker can become an entrepreneur merely if he wants to j 
or if every member of the community has to be mayor at least once — 
then we encounter a type and a degree of intrageneration m.obility 
that makes class formation and class conflict impossible. In this case, 
class belongingness becomes an accidental or merely temporary occur- 
rence. Although there still is a quasi-group structure of authority 
roles, the continuous exchange of their incumbents makes impossible 
the organization of interest groups defending or attacking the legiti- 
macy of authority structures: there is no class conflict, and there are 
no classes in the strict sense. 

As a mobile society (of the intrageneration variety), the classless 
society is thus a sociological category of realistic significance. How- 
ever, one qualification to this conclusion is necessary. There is some- 
thing to be said for an empirical generalization which Mosca calls the 
"law of inertia": "All political forces have the property which in 
physics is called inertia, i.e., a tendency to stay in a given state" (24, 
p. 61). Ben-David has specified this "law" with respect to Israel. 
For him, classlessness in the indicated sense characterizes "revolu- 
tionary periods" of social development rather than lasting types of 
social order ( 125, p. 303). Usually, these periods last but a few years. 
Then, that "articulation of the power structure" and "functional dif- 
ferentiation" sets in which Ben-David demonstrates for the social 
structure of the professions in Israel (p. 309). Although the rulers 
of totalitarian states like to operate with an "ideology of perpetual 
national emergency," as demonstrated impressively by Bendix (138, 
p. 443), for which the well-known theory of the "permanent revo- 
lution" is an example, it is plausible that the indicated state of quasi- 
classlessness is never more than a combination of transitory processes 
of radical change (which soon gives way to a minimum of stability) 
and of monopolization of power that makes possible the formation 
of classes and conflict between them. Classlessness by (intragenera- 
tion) mobility is, in sociological analysis, a limiting case that always 
tends toward its own abolition and that may therefore be ignored. 
There is no reason to assume that a stable society can operate on the 

222 Conflict Groups, Group Conflicis, and Social Change 

principle of the continuous patterned exchange of the personnel of 
authority positions." 

While social mobility, apart from its limiting case of permanent 
exchange between the classes within generations, cannot thus be said 
to present an obstacle to the formation of classes and the existence of 
class conflict, there can be little doubt that it affects the intensity of 
class conflict. From the point of view of mobility two distinct types 
of classes may be distinguished. With respect to dominant conflict 
groups, Mosca calls one of these the "aristocratic class" bent on the 
"maintenance of authority for the descendants of those who possess 
it at a given point of time," and distinguishes it from the "democratic 
class" characterized by the "tendency to rejuvenate the ruling class 
by upward mobility of persons from the ruled class" (24, p. 322). 
G. D. H. Cole confines the term "class" to the first of these types and 
refers to the second as "elite" (39). In analogy to those terms of 
Max Weber's frequently used in connection with social mobility I 
would recommend speaking here of "closed" and "open" classes. 
Where allocation to authority positions is based on ascriptive criteria, 
we find closed classes. By contrast, open classes are recruited anew 
in every generation. These types are nevertheless but two points on 
a scale of numerous gradations. From caste-like rigidity to quasi- 
classlessness there is a continuum of types of social classes determined 
by degrees of inter- and intrageneration social mobility. It seems 
plausible that this continuum also defines a scale of conflict intensity. 
There is an inverse relation between the degree of openness of classes 
and the intensity of class conflict. The more upward and downward 
mobility there is in a society, the less comprehensive and fundamental 
are class conflicts likely to be," As mobility increases, group solidarity 
is increasingly replaced by competition between individuals, and the 
energies invested by individuals in class conflict decrease. 

It is easy to see that the correlations between conflict intensity and 
empirical variables suggested in the preceding sections of this chapter 

^° There is possibly some connection between the size of associations and the 
feasibility of the principle of alternating chairmanship: the larger the association, the 
smaller the probability that complete openness of the authority positions will be main- 
tained for longer periods of time. It may be noted, here, that even in the Soviet 
Union purges have become more difficult technically today than they were in 1935. 

^^ From this point of view, the limiting case of quasi-classlessness becomes part 
of the intensity scale; it is distinguished in this sense not by the absence of classes, but 
by an intensity of conflict amounting to zero. In view of our assumption that class 
conflicts are universal, this seems the most plausible formulation of the case. 

Conflict Groups J Group Conflicts y and Social Change 223 

all involve a psychological factor as well. The intensity of conflicts 
is a function of the involvement of individuals. Earlier I have sug- 
gested that this involvement is likely to be greater if the individual 
participates in specific conflicts with several of his roles than if he 
participates only with one. With respect to mobility, our proposition 
might be reformulated in psychological terms also. If the individual 
sees for his son, or even for himself, the chance of rising into the 
dominant or falling into the subjected class, he is not as likely to 
engage his whole personality in class conflicts as he is when class posi- 
tion is of a more permanent nature. While in general these psycho- 
logical assumptions are probably safe to make, it has to be recognized 
that from the individual's point of view nonstructural factors may also 
influence his involvement in group conflict. Without doubt there are 
psychological constellations that make one individual more "quarrel- 
some" than another. I would suggest that individual variations of 
this kind are of but minor significance for the over-all intensity of 
class conflict j at the same time, their presence must not be overlooked 
in a comprehensive analysis. 


All the factors discussed so far have been related primarily to 
variations in the intensity of class conflict. Their effect on the violence 
of conflict manifestations seems but slight. The reverse is true with 
respect to the final, and in many ways most crucial, factor affecting 
the empirical patterns of class conflict: conflict regulation. Probably, 
eflFective conflict regulation is also of some consequence for the in- 
tensity of group conflict j but this important process is above all con- 
cerned with expressions or manifestations of conflict, and it therefore 
determines degrees of violence rather than intensity. In discussing 
the regulation of social conflict I shall start with a number of seem- 
ingly terminological problems. As the patient reader will soon realize, 
however, every one of these is tied to substantial problems of patterns 
of conflict, so that the distinction between terminological and substan- 
tial discussions becomes largely spurious. 

There are a number of competing concepts for what we shall call 
conflict regulation, some of which have connotations not intended here. 
This is especially true with respect to the concept of conflict resolution. 
The idea that conflicts may be resolved could mean, and often is 
thought to imply, that it is possible to eliminate given conflicts alto- 
gether. As used by some, the notion of conflict resolution addresses 
itself to the causes rather than the expressions of social conflict. This 

224 Conflict Groups, Group Conflicts , and Social Change 

notion is at the very least misleading, and I propose to reject it on 
theoretical grounds. There is one, and only one, sense in which one 
might say that a conflict has been resolved: the specific issues of a 
specific conflict, e.g., the claim of a union for a certain wage increase 
on a certain date, may be settled in such a way as not to reappear again. 
But it should by now be abundantly evident that from the point of 
view here suggested such specific settlements do not in the least 
affect the causes and determinants of even the specific conflict of the 
example j if, with changing issues, this conflict persists it has not there- 
fore been resolved. The concept of conflict resolution will be rejected 
as reflecting a sociologically mistaken ideology according to which 
complete elimination of conflict is possible and desirable." 

A second notion that has to be rejected as sociologically meaning- 
less is that of the suppression of social conflict. It is perhaps evident 
without discussion that suppression cannot be thought of as an effec- 
tive (or indeed desirable) means of regulating social conflicts. How- 
ever, I would go one step further and assert that effective suppression 
of conflict is in the long run impossible. This assertion contains the 
invariably awkward expression "in the long run." Even in our own 
century, history has shown that for totalitarian regimes it is possible 
to suppress opposition both in industry and in the state for what are, 
for the people concerned, very considerable periods of time. By using 
the phrase "in the long run," I do not want to be misunderstood as 
taking a cynical view of facts of this kind. It seems to me, however, 
that the very history of contemporary totalitarian states has demon- 
strated that the "long run" in this case extends to no more than a 
decade at the most. To the superficial observer this may seepi a sur- 
prising statement J unfortunately, I can here substantiate it only by 
one or two rather gross assertions. Where an attempt is made to 
suppress conflict altogether, either of two consequences is likely to 
occur within at most a decade. Either suppression amounts to com- 
plete nonrecognition and exclusion of opposition, in which case revo- 
lutionary changes of the Hungarian type are virtually bound to occur j 
or suppression of opposition is coupled with a careful and continuous 
scrutiny of the embryonic manifest interests of the potential oppo- 
sition, and changes are introduced from time to time which incorpo- 

^^ I recognize of course that the term "conflict resolution" might be defined dif- 
ferently. Thus, I have no doubt that the editors of the journal Conflict Resolution are 
really concerned with conflict regulation. However, it seems to me necessary to empha- 
size in the strongest possible terms the assumption of the universality of conflict — 
regulated or not. 

Conflict Groups, Group Conflicts y and Social Change 225 

rate some of these interests/^ In the latter case, suppression is not 
complete, and violent conflicts may simmer under the surface for a 
long time before they erupt j the ineflFectiveness of the former type 
of suppression needs no comment. 

We are here concerned not with resolution and suppression but 
with the regulation of social conflict. By this we mean such forms of 
conflict control as address themselves to the expressions of conflicts 
rather than their causes, and as imply the continued existence of an- 
tagonisms of interest and interest groups. Effective conflict regula- 
tion in this sense presupposes the presence of at least three factors, 
each of which in itself influences the violence of conflict manifes- 

First, for effective conflict regulation to be possible, both parties 
to a conflict have to recognize the necessity and reality of the conflict 
situation and, in this sense, the fundamental justice of the cause of 
the opponent. In a way, this is a value premise. Recognizing the 
justice of one*s opponent's cause does not mean of course that the 
substance of the opponent's interests has to be recognized as justified 
at the outset. Rather, recognition means, here, that both parties 
accept their conflict for what it is, namely, an inevitable outgrowth 
of the authority structure of associations. Wherever the attempt is 
made to dispute the case of the opponent by calling it "unrealistic," 
or denying the opponent the right to make a case at all, effective regu- 
lation is not possible. This is also true, however, where conflicts are 
not recognized for what they are, and where too great an emphasis 
is put on what are often misleadingly called "common interests," It 
seems to me that the London Economist was well-advised when it 
"reproached British unions for their ^moderation' which it declared 
in part responsible for the stagnation and low productivity of British 
capitalism J it compared their policies unfavourably with the more 
aggressive policies of American unions whose constant pressure for 
higher wages has kept the American economy dynamic" (see 80, p. 
198). Without doubt, there are "common interests" in any conflict 
situation; without community, no conflict, and vice versa. However, 
the crucial factor for effectively regulating conflicts is recognition, 
and even emphasis, of systematic divergence and opposition. The 
attempt to obliterate lines of conflict by ready ideologies of harmony 

^^ This "subtle totalitarianism," as one might call it, can be amply illustrated by 
the history of the Soviet Union. The most striking example is perhaps provided by 
the history of land collectivization (under Stalin), where the peasants' interests were 
increasingly recognized by the government. 

226 Conflict Groups, Group Conflicts , and Social Change 

and unity in effect serves to increase rather than decrease the violence 
of conflict manifestations.^* 

A second prerequisite of effective conflict regulation is the organi- 
zation of interest groups. So long as conflicting forces are diffuse, in- 
coherent aggregates, regulation is virtually impossible. Here, Coser's 
(or Simmel's) point is quite pertinent, namely, that despite the para- 
dox of the situation, conflict groups are often actually intent on fos- 
tering each other's unity and organization. "A unified party prefers 
a unified opponent" (8i, 132). Guerrilla warfare is not susceptible 
of effective regulation.^"^ Dubin believes that it is feasible to make a 
general proposition to the effect that "conflict between groups becomes 
institutionalized" (77, p. 287). Once again, however, the problem 
of the time span (covered by Dubin's "becomes") arises. Possibly, 
the inertia principle applies to this aspect of group conflicts also, so 
that "in due course" most conflict groups come to be organized. We 
have seen, however, that this organization depends above all on cer- 
tain structural conditions which are not universally present. From the 
point of view of effective conflict regulation, these conditions of or- 
ganization may be said to be one of its prerequisites, since the organi- 
zation of conflict groups itself is one of its prerequisites. 

Thirdly, in order for effective regulation to be possible, the op- 
posing parties in social conflicts have to agree on certain formal rules 
of the game that provide the framework of their relations. Again, 
Dubin seems to think that such agreement invariably comes about: 
"Continuous conflict between groups leads to standardized modes of 
conflict. . . . Continuous conflict between groups leads to routinized 
interactions" (77, p. 190). Here, too, however, Dubin's propositions 
seem overly optimistic. By rules of the game we shall understand 
such procedural norms as are binding for the contestants without prej- 
udicing the outcome of the contest. Normally, they would include 
stipulations as to where and how to meet, how to proceed, how to reach 
decisions, what sanctions to apply in case of noncompliance, and when 
and how to change the rules themselves. Kerr has pointed out that 
such rules are generally advantageous for conflict regulation as well 
as for the interest groups involved: "These rules normally protect 
the survival of both parties, reduce the potential injury to each, intro- 

^* One illustration of such ill-conceived attitudes and their consequences will be 
given in the discussion of the German codetermination experiment in Chapter VII. 

^^ I am sure that this point will be confirmed emphatically by all Colonial Secre- 
taries of mid-twentieth century governments. Very often, it is the absence of an 
organized opponent that makes the regulation of colonial disputes in the contemporary 
world so difficult. 

Conflict Groups y Group Conflicts y and Social Change ii'] 

duce some predictability into their actions, and protect third parties 
from undue harm" (84, p. 235). Examples for the establishment of 
such rules are numerous in industrial and political as well as interna- 
tional conflict. It is important to note, however, that rules of the game 
can serve their function only if and as long as they put both parties on 
an equal footing and do not imply any substantive stipulations dis- 
abling one or the other conflict group." 

Once these prerequisites of conflict regulation are present, vary- 
ing forms of regulation itself can come into operation. While a com- 
prehensive survey of such forms is neither possible nor necessary here, 
it seems useful to examine some of the more frequent types of con- 
flict regulation in terms of their eff^ect on the violence of class con- 
flict. In this field, generalization is of course as difficult as it is un- 
satisfactory. The conditions of different associations differ greatly, 
and analogies between types of conflict regulation in, say, industry and 
the state, are bound to do injustice to the specific requirements of either 
of these associations. Moreover, there are significant differences even 
between societies of similar forms of government; no generalization 
is feasible, e.g., about the place of the legal institutions of the United 
States, France, and Germany, in the regulation of political and in- 
dustrial conflicts. For these and similar reasons, I shall confine myself 
In the following discussion of forms of conflict regulation to a few 
highly abstract remarks which require both extension and modifica- 
tion with respect to specific associations and societies." 

^® The rules of the game are, from this point of view, always precarious. What Is 
a formal agreement at one point may turn out to prejudice the case of one of the op- 
ponents at a later point. For this reason, rules for changing the rules are perhaps the 
most important element of rules of the game (such as, e.g., constitutions). 

^^ To facilitate the following discussion, it may be useful to present a schematic 
picture of forms of conflict regulation. The following — highly tentative — schema rep- 
resents an extension of a table given by Moore (157, p. 446) and includes a list of 
terms suggested by Kerr (84, p. 236). It is evidently derived from conditions of 
conflict regulation in industry, and its adaptation to other associations presents certain 
difficulties. The schema is based on the role of third parties in conflict regulation. For 
further discussion of the types of regulation see the text. 

Role of Third Party in the Regulation of Social Conflict 

Invitation of 

Acceptance of 


3rd Party Advice 

3rd Party Advice 

Kerr's Terms 
















>■ arbitration 




J (suppression) 

228 Conflict Groups, Group Conflicts, and Social Change 

The first and foremost form of conflict regulation appears to con- 
sist in the operation of certain institutions which provide the frame- 
work for discussions of and decisions about conflicting issues. In gen- 
eral these institutions may be described as parliamentary or quasi- 
parliamentary bodies in which conflicting interest groups or their rep- 
resentatives meet in order to carry on their conflict in a relatively peace- 
ful and patterned manner. For such institutions to be effective, they 
have to comply with at least four standards: they must be autono- 
mous bodies invested with the right to reach decisions without having 
recourse to outside agencies of any kind j their place in a given associa- 
tion has to be monopolistic in the sense that they are the only institu- 
tion of its kind; their role must be obligatory both in that conflicting 
interest groups have to refer to these institutions in case of acute con- 
flicts, and in that decisions reached in them are binding for both interest 
groups and their members; and they have to be democratic, i.e., both 
parties have to be heard and be given a chance to state their claims 
before decisions are reached. Further procedural arrangements as to 
the modes of discussion and decision are part of the rules of the game. 

It seems evident that the very creation of parliamentary or quasi- 
parliamentary bodies of this kind involves a considerable reduction in 
the violence of group conflicts. Those who have agreed to carry on 
their disagreements by means of discussion do not usually engage in 
physical violence. Moreover, the violence of conflicts would seem to 
decrease as the efi^ectiveness of parliamentary institutions increases. 
At the same time, the presence of such institutions does not generally 
guarantee that violent conflicts will be avoided altogether.^^ In many 
cases, in order for the violence of group conflict to be effectively re- 
duced, autonomous conciliation has to be supplemented by other forms 
of conflict regulation. These other forms differ from conciliation in 
that all of them involve the intervention of "third parties,** i.e., out- 
side agencies. 

The mildest form of outside interference with group conflicts 
would seem to consist in what in industry is usually called mediation. 
Here, both parties agree to consult an outsider who is asked to give 

^^ The feasibility of majority decisions and, underlying this, the relative per- 
manence of authority structures are relevant here. In the association of the state where 
periodical elections are part of the rules of the game, majority decisions are an accept- 
able norm, although they may put one of the parties at a permanent disadvantage during 
one election period. In industry, majority decisions are not feasible, because the 
authority structure of the enterprise is not subject to changes by elections. The likeli- 
hood of breakdowns of autonomous conciliation is therefore considerably greater. 

Conflict GrowpSj Growp Conflicts y and Social Change 229 

advice but whose advice will have no binding force for the parties. At 
first sight, this type of regulation may seem to promise little effect; 
yet experience in many spheres of social life has shown that mediation 
in fact often is the most successful type of conflict regulation.^" Kerr's 
excellent analysis of mediation suggests that this kind of third party 
interference generally has at least five favorable consequences for con- 
flict regulation: reduction of irrationality, removal of nonrationality, 
exploration of solutions, assistance in graceful retreat, and raising the 
cost of conflict (see 84, pp. 236-39). Examples of this form of con- 
flict regulation are frequent in industrial and international conflicts, 
but rather rare in political conflict, although in some cases either cer- 
tain legal institutions or the incumbents of highly representative posi- 
tions (king, president) may serve as mediators between political par- 

The place of legal institutions in conflict regulation is more ac- 
curately described, however, by what in industry is usually called ar- 
bitration. There are two types of arbitration, differing in the kind of 
commitment to an arbitrator to which conflicting parties have agreed. 
The rules of the game may stipulate either that conflicting parties are 
obliged to call in an arbitrator in case of breakdown of conciliation and 
mediation but are free to accept or reject the arbitrator's decision, or 
that they are free to call in an arbitrator but must accept his decision 
once they have done so. In either case, arbitration is likely to be an 
effective means of reducing the violence of class conflict, although ar- 
bitration in general can, as Lockwood has pointed out (86), become 
a problematical form of conflict regulation. There is, according to 
Lockwood, a "political" and a "judicial" conception of arbitration. 
The first suggests that it is the task of arbitration to find a workable 
compromise between conflicting issues accepted as such; in terms of 
conflict theory, this approach promises success. The second conception, 
however, views conflicts from a legalistic point of view, i.e., ascribes to 
the arbitrator the task of judging the merits of conflicting issues in 
terms of fixed standards of "right" and "wrong." Where this is the 
case, the rules of the game are likely to prejudice the case of one or 
the other party, for if one side or claim is declared right in an ultimate 
legal and moral sense, conflict itself is not recognized and the other 
party is likely to feel so frustrated as to resort to violence. From in- 

^® Incidentally, this holds not only in social life, but also in psychological matters. 
Many psychoanalysts interpret their role as that of a mediator of conflicting emotions 
within their patients. 

230 Conflict Groups, Group Conflicts, and Social Change 

dustry, this dichotomy may be transposed to the political and legal 
systems. Thus, a comparative analysis of the legal systems of Anglo- 
Saxon and Continental European countries would probably show that 
whereas in the former a political conception of legal arbitration pre- 
vails, the latter display the judicial conception to an extent which is 
dangerous from the point of view of effective conflict regulation.^" 

The three forms of conflict regulation thus indicated, i.e., concilia- 
tion, mediation, and arbitration, may operate as successive stages of 
conflict regulation or be applied individually in given situations. If 
they are understood as a succession of stages, their logical conclusion 
would appear to be that type of arbitration in which both the invitation 
of an arbitrator and the acceptance of his decision are compulsory for 
the parties involved. However, compulsory settlement of disputes is 
not, in the sense of conflict theory, an effective mode of conflict regu- 
lation. Even if agreed upon by conflicting parties as a rule of inter- 
action, it seriously restricts their chances of defending their respective 
cases. There is, in compulsory settlement, at least a danger of one or 
the other party being dominated by an outside agency, e.g., govern- 
ment. In other words, compulsory settlement may lead to the sup- 
pression of conflict, with all the consequences indicated earlier.^^ 

There is an immense variety of empirical modes of conflict regu- 
lation 5 but I suggest that most of these concrete forms represent some 
modification or combination of the general types outlined in this sec- 
tion. Conciliation, mediation, and arbitration, and their normative and 
structural prerequisites, are the outstanding mechanisms for reducing 
the violence of class conflict. Where these routines of relationship are 
established, group conflict loses its sting and becomes an institutional- 
ized pattern of social life. For revolutionary upheavals to be trans- 
formed into evolutionary changes, there is, contrary to Marx's belief, 
no need for a classless society (that is, for a Utopian fiction) ; by effec- 
tive regulation, class conflict may become the element of regularity in 
a continuously changing world. Even if the intensity of conflict re- 
mains undiminished its manifestations may be channeled in such a way 
as to protect the individual from the physical threat of a bellum 
omnium contra omnes. In this sense, conflict regulation seems a more 

^^ Short of a comprehensive study, this point could be brought out, I think, by 
a detailed analysis of the position of defense counsels in Anglo-Saxon and Continental 
European courts of law, 

^^ This conclusion is borne out by two prominent experiences in the compulsory 
settlement of industrial disputes: that of Germany in the 1920's and that of con- 
temporary Communist states. 

Conflict Groups, Group Conflicts, and Social Change 231 

satisfactory solution, both theoretical and political, of the "Hobbesian 
problem of order" than the solution offered by Talcott Parsons.^^ 


Throughout the abstract considerations of the preceding chapters, 
we have assumed that group conflicts of the class type result in struc- 
ture changes. We have moreover rejected attempts to distinguish be- 
tween allegedly different kinds of change such as "changes within" 
and "changes of" the structure of associations. Having now assembled 
the elements of a theory of class conflict, it is our task, in conclusion, to 
specify how structure changes are brought about by class conflict and 
under which specific conditions particular modes of structure change 
must be expected. Again, this involves supplementation of a model 
by terminological distinctions and empirical generalizations. 

Like the stability of social structure, changes in class structure 
must be investigated on the two levels of analysis which we have called 
the normative, or ideological, and the factual, or institutional, levels. 
Interests may become values, but, also, realities. If, for example, one 
of the manifest interests of a conflict group consists in equality in the 
sense of Marshall's citizenship rights, there are two levels on which 
this interest may become realized. First, it may become a more or 
less general value orientation of the citizens of a society j Tocqueville's 
"manly and lawful passion for equality" may spread so "that [it] 
incites men to wish all to be powerful and honored" (232, p. s^)- 
Secondly, equality may be incorporated in institutional arrangements, 
as by the socialization of medicine, the abolition of school and uni- 
versity fees, etc. Both processes are processes of change, and it would 
be a uselessly speculative undertaking to try and establish a primacy 
of either the normative or the institutional level. While, therefore, 
neither of these levels must be neglected in any analysis of structure 
change, it nevertheless seems necessary to use an operational approach 
that specifies modes of structure change in terms of but one of these 
levels. In the present context, all structure changes will be understood 
as changes involving the personnel of positions of domination in im- 
peratively coordinated associations. 

^^ The problem of conflict regulation constitutes the most important consequence 
of conflict theory in terms of social policy. I have only intimated this aspect of the prob- 
lem in the preceding discussion, and its conclusions should not be regarded as auto- 
matically applicable to the political problem of conflict regulation. Thus, violent con- 
flict may at times actually be desirable in some associations. Generally speaking, how- 
ever, it would seem to be the task of social policy to try to regulate the inevitable con- 
flicts of social life by means other than resolution or suppression. 

232 Conßict Groups, Group Conflicts, and Social Change 

The immediate effect of class conflicts is brought to bear on the 
incumbents of positions of domination, and changes are introduced by 
way of these positions. 

This operational specification seems tenable in view of the role and 
definition of authority in the theory here proposed. Insofar as the 
distribution of authority in associations can be described as the formal 
object of class conflict, changes resulting from class conflict are in their 
formal aspect always changes in these authority structures or in their 
personnel. Moreover, authority is, from the point of view of socio- 
logical analysis," an instrumental value. In class theory, the possession 
of authority does not figure as a value sought for its own sake but as 
an opportunity to realize specific interests. This conception of author- 
ity is in keeping with our distinction of power and authority by the 
central category of legitimacy. It follows from it that, e.g., an ex- 
change of the personnel of positions of domination has to be viewed 
not only as a process of rejuvenation of a basically constant "ruling 
class" or "elite," but above all as the instrumental aspect of a process 
which substantively represents structure change. In this sense, ex- 
changes of personnel are not in themselves structure changes, but 
merely a condition for (from the point of view of the status quo) 
"new" interests becoming values or realities. Problems of changing 
patterns of recruitment to an upper stratum are meaningful only in 
the context of integration theory. For coercion theory, changes of the 
personnel of authority roles are merely the formal or instrumental 
aspect of changes of social structure on both the normative and the in- 
stitutional levels. 

The operational approach suggested permits the distinction of at 
least three modes of structure change, each of which requires some 
comment. A first mode of change, in this sense, consists in the total 
(or near-total) exchange of the personnel of positions of domination 
in an association. This is clearly the most sudden type of structure 
change. For purposes of illustration we might use the association of 
the state, and assume a specific state to be divided in such a way that 
there are three political parties, two of which are in opposition and one 
in power. A total exchange of the personnel of positions of domination 
would involve, then, the replacement of all cabinet ministers, higher 
civil servants, and other incumbents of political ofiice by members of 

^^ As against, e.g., psychological analysis, where "lust for power" and the posses- 
sion of power are of considerable interest, also, as immediately gratifying values. 

Conflict Groups y Group Conflicts y and Social Change 233 

the opposition parties. In modern states, such sudden changes are a 
comparatively rare occurrence j the last outstanding example was prob- 
ably provided by the Bolshevik revolution in Russia. Generally speak- 
ing, total exchange of ruling personnel might also be described as revo- 
lutionary change. It is at this point that the sociology of revolution 
ties in with the theory of group conflict. 

Far more frequently we encounter in history, and especially in 
modern history, a second mode of structure change, namely, the par- 
tial exchange of the personnel of positions of domination. Such par- 
tial exchange signifies evolutionary rather than revolutionary change. 
In terms of our illustration, it would be present if the majority party 
chose to form a coalition with, say, the smaller one of the two opposi- 
tion parties. In this case, some representatives of the hitherto sub- 
jected class penetrate the ruling class and influence the policies adopted 
and decisions made. Coalitions are of course not the only example of 
partial exchange j if elections in democratic countries reverse the ma- 
jority relations, they usually result in but partial exchanges of gov- 
ernment personnel, so that, e.g., cabinet ministers are exchanged, but 
some of the judges, diplomats, and higher civil servants of the pre- 
vious majority party remain in office.^* 

But probably more important than either of these is a third mode 
of structure change by class conflict which does not involve any ex- 
change of personnel. It is possible for structure changes in directions 
intended by subjected groups to be inaugurated without any members 
of these subjected groups penetrating into dominant positions. This 
seemingly accidental consequence of the process of social conflict occurs 
in democratic and totalitarian countries alike. In terms of our example, 
it would mean that majority and opposition remain stable and dis- 
tinct over long periods of time, but the majority party incorporates 
proposals and interests of the opposition in its legislation and policies. 
Strange as it may initially appear that structure change should ever 
occur without an exchange of ruling personnel, there are nevertheless 
numerous illustrations in the history of states, enterprises, churches, 
and other associations. To be sure, this third mode of structure change 
marks the slowest type of evolution and requires particular skill on 

^* There are considerable diflPerences between different countries in this respect, 
ranging from the German case where judges and civil servants cannot be fired, to the 
American case where replacement of leading personnel tends to include even not-so- 
senior civil servants. 

234 Conflict Groups, Group Conflicts, and Social Change 

the part of the rulers to avoid such suppression of opposing interests 
as thereby to provoke revolts^' j it can nevertheless enable a dominant 
class to maintain the legitimacy of its authority over long periods of 

Possibly, these three modes of structure change indicate the end 
points and center of a scale that measures the suddenness of change. 
Partial exchange of personnel is evidently a broad category that covers 
the whole field between total exchange and complete stability. How- 
ever, while it may be said that structure change is more sudden in the 
extent to which more personnel is exchanged, this does not necessarily 
mean that it is also more radical. Suddenness and radicalness of struc- 
ture change are two dimensions of this phenomenon which can vary 
independently, much as the intensity and violence of class conflicts 
can vary independently. There are examples of relatively sudden 
changes that are accompanied by but slight modifications of values and 
institutions, and there are examples of extremely radical, although 
comparatively slow, evolutions.^® Majority shifts in democratic states 
illustrate the former case, while the latter is illustrated by such deep 
changes in class structure as we have analyzed earlier in this study. 

The relation between the radicalness-suddenness dimension of 
structure change and the intensity-violence dimension of class con- 
flict is more than merely logical. It may be argued that the suddenness 
of change varies directly with the violence of conflict. The more vio- 
lent class conflicts are, the more sudden are the changes wrought by it 
likely to be. In this sense, eff^ective conflict regulation serves to re- 
duce the suddenness of change. Well-regulated conflict is likely to 
lead to very gradual change, often near the third mode distinguished 
above. Conflict regulation may, in fact, constitute a machinery for 
forcing on dominant groups recognition of the interests of subjected 
groups, which interests are then incorporated in policy. The example 
of a wage claim settled by conciliation is a case in point. Uncontrolled 
conflict, on the other hand, always threatens the incumbents of posi- 

^^ The history of the Catholic church provides examples of both: the skiUful 
handling of slow changes of policy without exchange of personnel, and the degenera- 
tion of this pattern into suppression and consequent revolt. 

^* In fact, the two concepts of revolution so often interchanged in the literature 
may be differentiated in terms of the distinction between suddenness and radicalness. 
The Industrial Revolution was probably more radical than the French Revolution, yet 
it was not nearly as sudden. The term "revolution" is often used indiscriminately for 
both particularly radical and particularly sudden changes. I should prefer to use it in 
the latter sense only. 

Conflict Groups, Group Conflicts, and Social Change 235 

tions of domination in their very possession of authority} it aims at a 
total exchange of leading personnel, and, in this sense, at sudden 

At least in theory, there is also a scale of the radicalness of struc- 
ture change. However, an operational formulation of such a scale 
oflFers particular difficulties. In general, the radicalness of structure 
changes is evidently a function of what in particular historical situa- 
tions represents the status quo. In eighteenth-century Europe, the 
peaceful utilization of nuclear energy would certainly have made for 
extremely radical changes, both technical and social. In twentieth- 
century Europe, the same process, although still involving some 
change, has no really radical consequences but simply ties in with con- 
tinuing trends of rationalization, automation, etc. Similarly — and 
more immediately to the point here — changes resulting from conflicts 
within the association of the Catholic church are, in most countries of 
today, far less radical than they were at earlier times. Thus, the radi- 
calness of structure change is not merely a consequence of the inten- 
sity of class conflict. Within certain limits, however, this relation does 
obtain. The more strongly people are involved in given conflicts, the 
more far-reaching are their demands likely to be, and the more radical 
will be the changes resulting from this conflict, irrespective of the 
suddenness of such changes. Radicalness and suddenness of change, 
like intensity and violence of conflict, may coincide, but more often 
they diverge} and in any case their divergence presents more interest- 
ing problems of social analysis than their coincidence. 

Apart from historical conditions, the co-variance of the intensity of 
conflict and radicalness of change as well as of the violence of conflict 
and suddenness of change is further restricted by the structural requi- 
sites of associations. This will become immediately apparent if we 
contrast industrial and political conflict. In political associations, ex- 
change of leadmg personnel is a realistic possibility, so that the whole 
scale of suddenness of change can be applied here. In industrial as- 
sociations, exchange of leading personnel is, except within certain 
very narrow limits, not possible." There are no elections (and there 
cannot be) as a result of which the members of management become 

''''These limits will be further explored in Chapter VII, below. In a formula: 
individual members of the subjected class can be taken into management, and the 
personnel of management can be totally replaced by revolutionary action. However, 
both these phenomena indicate structure changes of a special type, namely, changes that 
do not satisfy the interests of a subjected group. 

236 Conflict Groups, Group Conflicts, and Social Change 

workers, and representatives of labor become managers. Structure 
changes in industry will therefore almost invariably take the form 
of changes of policy unaccompanied by exchanges of personnel. To 
some extent, similar conditions obtain in most church organizations. 
Before investigating processes of structure change by class conflict in 
specific associations, we must therefore ascertain the range of pos- 
sible modes of change in these associations."^ 

In general, however, and without losing sight of additional inter- 
vening variables, we can propose that different modes of structure 
change co-vary with different modes of class conflict. The more in- 
tense class conflict is, the more radical are the changes likely to be 
which it brings about j the more violent class conflict is, the more sud- 
den are structure changes resulting from it likely to be. Structure 
change is the final element of the theory of group conflict under dis- 
cussion. Like all other elements of this theory, it represents but a 
segment of more inclusive phenomena of conflict and change. Pos- 
sibly, the typology of change introduced in the preceding pages is 
applicable also to the consequences of kinds of conflict other than that 
between classes. As everywhere in this study, however, I have con- 
fined myself here to exploring the causes, forms, and consequences of 
conflicts generated by the authority structures of imperatively co- 
ordinated associations. 


It may appear premature, if not overambitious, to have used the 
word "theory" in connection with the approach outlined in the last two 
chapters. I have suggested a number of premises, concepts, models, 
and empirical generalizations which appear to have a bearing on prob- 
lems of social conflict and social change, but these suggestions do not 
display a degree of formalization and rigidity that would warrant 
calling them a theory. I may say that this shortcoming was partly de- 
liberate, partly unavoidable. It was deliberate in that attempts at for- 
malization in sociology have, in my opinion, remained to the present 
day more pretentious than useful, and that I understandably wanted to 
avoid seeing the approach suggested in this study exposed to the same 
criticism. At the same time, I would admit that formalization in so- 

^® The specific conditions of industrial associations are clearly related to the fact 
that whereas it is possible to conceive of modes of regulation which exclude civil war 
forever, strikes cannot be completely avoided. There is, so to speak, no outlet for in- 
tense conflicts within the social structure of industry, so that intensity is sometimes 
transformed into violence. 

Confiict Groups y Group Conflicts, and Social Change 237 

ciology is desirable. From this point of view, I regret not to have been 
able to give a more rigid formulation to the approach discussed in the 
last two chapters. If in conclusion of this abstract analysis I shall try 
now to summarize the main points of my approach to group conflict, 
the seemingly systematic character of this summary should not be mis- 
taken as a statement that complies with the methodological standards 
of scientific theories. Rather, I hope that this summary may enable 
other students in the field to advance beyond the limits of my own 
capabilities of formalization. 

1. The approach of this study has to be understood in terms of 
two premises — one formal, one substantive — which, although they 
are of a meta-theoretical or methodological nature, provide the neces- 
sary frame of reference of its elements. 

1.1. The heuristic purpose of the approach proposed in the pres- 
ent study is the explanation of structure changes in terms of group 
conflict. This purpose is therefore neither purely descriptive nor re- 
lated to problems of integration and coherence in or of society. 

1.2. In order to do justice to this heuristic purpose it is necessary 
to visualize society in terms of the coercion theory of social structure, 
i.e., change and conflict have to be assumed as ubiquitous, all elements 
of social structure have to be related to instability and change, and 
unity and coherence have to be understood as resulting from coercion 
and constraint. 

2. Within this frame of reference, the theory of social classes and 
class conflict involves a number of concepts to be defined. 

2.1. ^^ Authority is the probability that a command with a given 
specific content will be obeyed by a given group of persons" (M. 
Weber, 33^, p. 152). 

2. 1. 1. By domination shall be understood the possession of au- 
thority, i.e., the right to issue authoritative commands. 

2.1.2. By subjection shall be understood the exclusion from au- 
thority, i.e., the duty to obey authoritative commands. 

2.2. "An association shall be called imperatively coordinated as- 
sociation insofar as its members are, by virtue of a prevailing order, 
subject to authority relations" (M. Weber, 33«, p. 153). 

2.3. Orientations of behavior which are inherent in social posi- 
tions without necessarily being conscious to their incumbents (role ex- 
pectations), and which oppose two aggregates of positions in any 
imperatively coordinated association, shall be called latent interests. 

2.4. Quasi-group shall mean any collectivity of individuals shar- 

238 Conflict Groups, Group Conflicts, and Social Change 

ing positions with identical latent interests without having organized 
themselves as such. 

2.5. Manifest interests shall mean orientations of behavior which 
are articulate and conscious to individuals, and which oppose collec- 
tivities of individuals in any imperatively coordinated association. 

2.6. Interest group shall mean any organized collectivity of in- 
dividuals sharing manifest interests. 

2.7. By social class shall be understood such organized or unor- 
ganized collectivities of individuals as share manifest or latent in- 
terests arising from and related to the authority structure of impera- 
tively coordinated associations. It follows from the definitions of 
latent and manifest interests that social classes are always conflict 

2.8. Any antagonistic relationship between organized collectivi- 
ties of individuals that can be explained in terms of patterns of social 
structure (and is not, therefore, sociologically random) shall be called 
group conflict. 

2.9. Class conflict shall mean any group conflict that arises from 
and is related to the authority structure of imperatively coordinated 

2.10. Any deviation of the values (normative structure) or in- 
stitutions (factual structure) of a unit of social analysis at a given 
point of time {T -\- n) from those of a preceding point of time {T) 
shall be called structure change, insofar as it involves the incumbents 
of positions of domination. 

2.10. 1. By radicalness of structure change shall be understood 
the significance of consequences and ramifications of structure change. 

2.10.2. By suddenness of structure change shall be understood 
the extent to which incumbents of positions of domination are re- 

3. The formation of conflict groups of the class type follows a 
pattern that can be described in terms of a model involving the fol- 
lowing partly analytical, partly hypothetical steps: 

3.1. In any imperatively coordinated association, two, and only 
two, aggregates of positions may be distinguished, i.e., positions of 
domination and positions of subjection. 

3.2. Each of these aggregates is characterized by common latent 
interests J the collectivities of individuals corresponding to them con- 
stitute quasi-groups. 

3.3. Latent interests are articulated into manifest interestsj and 

Conflict Groups, Group Conflicts , and Social Change 239 

the quasi-groups become the recruiting fields of organized interest 
groups of the class type. 

3.3.1. Articulation of manifest interests and organization of in- 
terest groups can be prevented by the intervention of empirically 
variable conditions of organization. 

3.3.2. Among the conditions of organization, technical conditions 
(personnel, charter), political conditions (freedom of coalition), and 
social conditions (communication, patterned recruitment) can be dis- 
tinguished. To these, certain nonstructural psychological conditions 
(internalization of role interests) may be added. 

4. The course of group conflict of the class type also follows a 
pattern that can be described in terms of a model involving both 
analytical and hypothetical elements. 

4.1. Once the formation of conflict groups of the class type is 
complete, they stand, within given associations, in a relation of group 
conflict (class conflict). 

4.1.1. The intensity of class conflict varies on a scale (from o 
to I ) according to the operation of certain factors. 

4. 1 . 1 . 1 . The intensity of class conflict decreases to the extent that 
the conditions of class organization are present. 

4. 1. 1.2. The intensity of class conflict decreases to the extent that 
class conflicts in different associations are dissociated (and not super- 
imposed) . 

4. 1. 1. 3. The intensity of class conflict decreases to the extent that 
different group conflicts in the same society are dissociated (and not 

4. 1. 1. 4. The intensity of class conflict decreases to the extent that 
the distribution of authority and the distribution of rewards and 
facilities in an association are dissociated (and not superimposed). 

4. 1. 1. 5. The intensity of class conflict decreases to the extent that 
classes are open (and not closed). 

4.1.2. The violence of class conflict varies on a scale (from o 
to I ) according to the operation of certain factors. 

4. 1. 2. 1. The violence of class conflict decreases to the extent that 
the conditions of class organization are present. The violence of class conflict decreases if absolute depri- 
vation of rewards and facilities on the part of a subjected class gives 
way to relative deprivation. The violence of class conflict decreases to the extent that 
class conflict is effectively regulated. 

240 Conflict Groups, Group Conflicts, and Social Change 

4.2. Group conflict of the class type eflfects structure changes in 
the associations in which it occurs. 

4.2.1. The radicalness of structure change co-varies with the in- 
tensity of class conflict. 

4.2.2. The suddenness of structure change co-varies with the vio- 
lence of class conflict. 

I must leave it to the reader to judge whether the preceding list 
of abstract statements provides a satisfactory summary of the ap- 
proach suggested in the last two chapters of this study. Obviously, 
both the summary and its more elaborate basis are, as they stand, 
tentative and in need of many a refinement. I would nevertheless 
claim that even in this preliminary and tentative form the approach 
to group conflict proposed here provides a useful tool of analysis. 
It directs our attention to a specific set of problems and furnishes these 
with a coherent explanation. It is, in this sense, a searchlight that 
illuminates one sector of reality. In the following chapters I shall 
try to substantiate this claim by applying the theory of social class 
and class conflict to some of the significant problems of post-capitalist 


Classes in Post-Capitalist Society 
I: Industrial Conflict 


Formulation and application of a theory are two different mat- 
ters, each of which obeys its own laws and patterns. While the theory 
itself can be set out in a highly schematic and "logical" fashion, the 
analysis of facts would lose much of its color and interest if forced 
into the strait jacket of theoretical exposition. Although I shall in- 
dicate when the following analysis of conflict in advanced industrial 
society is guided by the theory of social class and class conflict, I shall 
not attempt to rearrange facts so as to fit the order of postulates, 
models, and hypotheses resulting from the considerations of the last 
two chapters. The order of reality rather than of theory will guide 
our analysis in the final chapters of this study, except in this section 
of this chapter, which serves a special purpose in the context of the 
following analysis. 

It is proper to demand that if we dismiss an old theory — as we did 
Marx's — and replace it with a new one, the new theory should be 
capable of explaining both the facts accounted for and the facts left 
unexplained by the old theory. Thus, one of the tests of the useful- 
ness of our theory of group conflict lies in its applicability to the con- 
ditions with which Marx dealt. There is, of course, no intrinsic reason 
why it should be possible to deal more schematically with this his- 
torical material than with post-capitalist society. However, in the 
present context I propose to simplify the task of reconsidering class 
conflict under capitalism. I shall refrain from questioning the facts 
described by Marx, and, instead, concentrate on how these facts appear 
in the light of the theory of group conflict. This would appear to be 
a doubtful procedure. All too often the societies that appear in the 
work of sociologists are merely historical constructions borrowed from 
earlier works or even invented in order to provide an impressive 

242 Classes in Post-Capitalist Society (I) 

contrast for contemporary data. Were it not for the thorough docu- 
mentation of Marx's work, and for the deliberate sketchiness of this 
initial section, I should do everything to avoid the suspicion of having 
an uncritical attitude toward history. As it is, I must ask the reader's 
indulgence if the next few pages leave much to be desired in terms of 
historical accuracy and detail. 

The starting point of Marx's analyses consists in what he himself 
variously calls the "sphere of production," the "relations of produc- 
tion," or "property relations." Clearly, all these expressions refer to 
the industrial enterprise and the social relations obtaining within it. 
For Marx, the enterprise is the nucleus of class war. In terms of our 
approach, the relevant feature here is that the industrial enterprise is 
an imperatively coordinated association. Marx, of course, emphasized 
the property aspect. This seems reasonable, in retrospect, since at his 
time it was legal possession of the means of production that provided 
both the foundation of capitalist power and the main issue of industrial 
conflict J but this is nevertheless too specific an approach to the problem. 
Industrial enterprise, being an imperatively coordinated association, 
has in it two quasi-groups which we may designate, following Marx, 
as those of capital or the capitalists and of wage labor or the wage 
laborers. Both capital and labor were united by certain latent interests 
which, being contradictory, placed them on the opposite sides of a 
conflict relation. While the most formal objective of the opposing 
interests was, in capitalist society, either the maintenance or the change 
of the status quo of authority, the precise substance of the conflict 
might, in relation to the specific conditions of this period, be described 
as a clash between capital's profit orientation and labor's orientation 
toward an improvement of their material status. 

The intensity of conflict in capitalist society was increased by the 
superimposition of authority and other factors of social status, espe- 
cially income. Domination meant, for the capitalists, a high income, 
while subjection involved for labor extreme material hardship. There 
was a clear correlation between the distribution of authority and social 

Despite this initial position, large obstacles were in the way of or- 
ganization for both quasi-groups in the early stages of industrializa- 
tion. We find here that constellation of factors described above (p. 
188) which makes the organization of interest groups virtually im- 
possible. Lack of leaders and ideologies (technical conditions), het- 
erogeneous modes of recruitment to authority positions (social condi- 
tions), and, in the case of labor, the absence of freedom of coalition 

Classes in Post-Cafitalist Society (I) 243 

(political conditions) — all these hold industrial conflict for some con- 
siderable time in a stage of latency, in which there are only occasional 
attempts at organization. As industrial associations stabilize, the con- 
ditions of organization gradually emerge, and both capital and labor 
form organizations (employers' associations, trade unions) in defense 
of what are now articulate manifest interests. Industrial class conflict 
enters a manifest phase of which strikes and lockouts are the most tell- 
ing symptoms. 

The situation described so far is that of the sphere of industry. It 
is characteristic of conflict in capitalist societies, however, that not only 
authority and social status, but also industrial and political conflict are 
superimposed one on the other. ^ The dominating groups of industry 
were at the same time the dominating groups of the state, either in 
person, through members of their families, or by other agents. Con- 
versely, the subjected groups of industry were as such excluded from 
political authority. Industry is the dominating order of society j its 
structures of authority and patterns of conflict therefore extend to the 
whole society. Consequently, the quasi-groups of industry also extend 
to the political sphere. The industrial quasi-group of capital becomes, 
as bourgeoisie (to use the Marxian terms once again), the dominant 
group of the state, whereas wage labor is, as proletariat, subjected in 
the political sphere as well. Since, under the particular conditions of 
capitalist society, conflict fronts that characterized industry and society 
were identical, the conflict was intensified to an extraordinary degree. 

In the political field, too, organization of conflict groups proved 
difficult in the beginning. Insofar as industrial and political quasi- 
groups were identical, the same factors were at work in the state that 
tended to prevent industrial organization. Moreover, political re- 
strictions, such as electoral systems, made it difficult for the proletariat 
to form effective interest groups. Thus, class conflict was smoldering 
below the surface of society for some time, until all restrictions fell 
and the two classes met openly in the political arena. 

By virtue of the superimposition of various lines of differentiation 
this conflict was, as we have seen, extremely intense. Its intensity was 
further increased by the fact that both classes were relatively closed 
units. Mobility within and between generations remained an excep- 

^ This is one of the points where detailed historical analysis would probablv come 
to correct the factual statements of Marx. While in English society in the nineteenth 
century industrial power may, indeed, almost automatically have conveyed political 
power, this is not true, e.g., for Germany, where, on the contrary, entrepreneurs were 
restricted in their exercise of industrial power b\- the impact of older political elites. 

244 Classes in Post-Ca-pitalist Society (I) 

tion.^ Bourgeoisie and proletariat were strictly separate and largely 
self-recruiting groups. But in this period it was not merely the in- 
tensity of the conflict but the violence as well that was extraordinarily 
great. In industry and the state, there were virtually no accepted 
modes of conflict regulation. In the absence of a democratic process 
that put both parties to a conflict on an equal footing, the subjected 
class increasingly became a suppressed class which faced as a solid but 
powerless bloc the absolute rule of the incumbents of roles of domina- 
tion. Because of this hardening of the class fronts, there were wide- 
spread demands for a complete and revolutionary change of existing 
structures. For structure changes could not slowly grow out of class 
conflict in this stage. Immobility and lack of regulation made the pen- 
etration of the ruling class by members of the subjected class impos- 
sible. At the same time, there were neither institutional channels nor 
ideological provisions for the ruling class to accept and realize any of 
the interests of the proletariat. Thus, it seemed justified to predict 
that class conflict in capitalist society tended toward both sudden 
and radical changes, i.e., a revolution promoted by the proletariat 
which replaces in one stroke the dominant groups of industry and 

Marx carried his analysis of capitalist society approximately to this 
point. Although he went considerably further in detail, his whole 
work converges on the prediction of the proletarian revolution. We 
have seen earlier how at this point Marx became a prisoner of pre- 
conceived philosophical and, perhaps, political convictions. Thus he 
did not, or would not, notice that factual developments followed the 
course of his predictions only up to a point. The ossification of conflict 
fronts and the intensification of conflict began to be checked both by 
the very fact of organization of interest groups on the part of the 
proletariat and by the structure changes to which this organization led. 
Within industry in particular, signs of the development of modes of 
regulation became apparent 5 trade unions managed to make some of 
their claims effectively heard and accepted. Marx showed himself a 
consistent philosopher but a poor sociologist when he tried to ridicule 
such "partial results" and the operation of trade unions (i.e., indus- 
trial conflict, as distinct from political class conflict) in general. His 

^ Here, again, the correctness of Marx's factual assertions is doubtful. It is hard 
to see how a new ruling class — capital — can be recruited without mobility, including 
mobility between the classes. It is obvious here that, even in the sense of our theory, 
Marx has geared his facts in such a way as to make plausible the extreme civil-war type 
of conflict which he anticipated. 

Classes in Post-Ca-pkalist Society (I) 245 

attempt to advocate, despite such tendencies, an intensification of the 
class war, and his insistence on the revolutionary goal of the prole- 
tariat, document his prophetic and political rather than his scientific 
self. At this point, we have to reject not only the substance, but the 
very intention of his work. 

Before we try to follow the indicated lines of class conflict some- 
what beyond the point of Marx's analysis, one clarifying remark 
seems in place. It should now be abundantly clear that the traditional, 
Marxian concept of class is but a special case of the concept advanced 
in the present study. For Marx, classes are conflict groups under con- 
ditions of {a) absence of mobility, {b) superimposition of authority, 
property, and general social status, {c) superimposition of industrial 
and political conflict, and {d) absence of effective conflict regulation. 
Thus, classes are conflict groups involved in extremely intense and 
violent conflicts directed toward equally extremely sudden and radi- 
cal changes. This is the "traditional" or "historical" concept of class. 
As against this concept, we have removed all four conditions men- 
tioned from its definition and included them as empirically variable 
factors in a theory of social class and class conflict. In this way, the 
concept itself becomes a highly formal and — in this sense — "unhis- 
torical" category j but the theory gains in fruitfulness, range, and ap- 

Thus, what has happened since Marx are in fact changes in the 
factors that contributed to the intensity and violence of the conflicts of 
his time. Patterns of conflict regulation emerged in both industry and 
the state. More and more, the democratic process of decision-making 
gave both parties a chance to realize their goals. The violence of class 
conflict was thereby effectively reduced. The institutionalization of 
social mobility made for a certain degree of openness in both classes. 
Absolute deprivation on the scales of social stratification gave way, for 
the proletariat, to relative deprivation, and later, for some, to com- 
parative gratification. Finally, the associations of industry and the 
state were dissociated to some extent. All these changes served to 
reduce both the intensity and the violence of class conflict in post- 
capitalist society, and to make sudden and radical structure changes in- 
creasingly improbable. New patterns of class conflict emerge, to which 
we shall turn presently. 

It must, of course, be emphasized that, whatever concept or theory 
one employs, history cannot be explained solely in terms of class. 
The changes that separate capitalist and post-capitalist society are not 
wholly due to the effects of class conflict, nor have they merely been 

246 Classes in Post-Ca-pkalisi Society (I) 

changes in the patterns of conflict. Thus, the subdivision of authority 
positions stimulated by an ideology of rationalization in both the en- 
terprise and the state is an autonomous process. The decomposition of 
capital and labor by the separation of ownership and control, and by 
the emergence of new differentiations of skill, has consequences for 
class conflict but is due to other factors. As a comprehensive process, 
the development from capitalist to post-capitalist society remains out- 
side the scope of the present analysis. But it should be clear from the 
preceding sketch that in principle our theory of group conflict is appli- 
cable, also, to the facts with which Marx dealt — and I hope it will be 
clear from the following rather more elaborate analysis in what sense 
it lends itself, by generalizing earlier approaches, to a coherent 
account of industrial and political conflict in the contemporary world. 


In a sense, Schelsky is undoubtedly right in calling the "often 
heard question . . .: Have we still got a class society today?" a 
"naive" question (72, p. 62). However, this question is naive not so 
much because it is too general to be answered with a plain "yes" or 
"no," but because it can be answered without thereby stating anything 
significant or exciting about post-capitalist society. Are there still 
classes? Or, as we can ask more precisely now: Are there still interest 
groups and quasi-groups in the sense of class theory? That there are 
interest groups in contemporary society can be affirmed immediately. 
There are, for example, trade unions and employers' associations, pro- 
gressive and conservative political parties. It is not difficult to show 
that all these organizations are interest groups in the sense of our 
definition. Quasi-groups, on the other hand, may be assumed to exist 
wherever there are authority relations and imperatively coordinated 
associations. Is it necessary to prove that there are such associations and 
relations in contemporary society? The state, the industrial enterprise, 
the churches — to mention only a few — are imperatively coordinated 
associations which exist in all modern societies and which, if our theory 
is right, justify the assumption that there are quasi-groups with con- 
flicting latent interests within them. And if post-capitalist society has 
quasi-groups and interest groups, it has classes also. Like its precursor, 
advanced industrial society is a class society. Concept and theory of 
class are still applicable. 

By taking this position we differ from a number of sociologists 
whose work has been discussed above. But is this difference, as de- 
scribed so far, more than a difference of terminology? Cannot the 

classes in Post-Cafitalist Society (I) 247 

charge be leveled against us that we presuppose the existence of classes 
by definition instead of demonstrating it empirically? Can we really 
answer the question of whether we still have a class society as easily 
as we did? 

The assertion that there still are classes because there are quasi- 
groups and interest groups is indeed less than a definition. It is, on the 
basis of class theory, a mere tautology. On the other hand, the asser- 
tion that there are still classes because there are imperatively coordi- 
nated associations is more than a definition. Although it presupposes 
the theoretical — and perhaps definitional — connection between classes 
and authority relations, it asserts the empirical presence of relations 
of authority. Social classes and class conflict are present wherever 
authority is distributed unequally over social positions. It may seem 
trivial to state that such unequal distribution exists in associations of 
post-capitalist society, but this assertion nevertheless establishes both 
the applicability of class theory and the radical difference from all at- 
tempts to describe contemporary society as classless. 

Nevertheless, to conclude merely that we are still living in a class 
society is as insufficient as it is unsatisfactory. It marks the beginning, 
not the end, of an analysis of advanced industrial society. For rhany 
people, the notion of a class society immediately evokes such definite 
associations that to apply it to a particular society might appear to in- 
volve a substantial statement of fact. I should like to emphasize there- 
fore that I do not regard it as such. I am concerned, here, not with 
asserting the applicability of class theory, but with applying it. If im- 
peratively coordinated associations can be shown to be a functional 
requisite of social structures, then the universal existence of classes is 
postulated by the same token. By way of empirical generalization we 
can maintain at the very least that in many societies there are associa- 
tions and classes, and in all known societies social conflicts. Societies do 
not differ by the fact that in some there are classes and in others not. 
Just as in the sociology of the family we are concerned not with the 
existence but with the patterns and functions of the family, so here we 
are dealing not with the presence of classes but with their nature and 
effect. By confronting capitalist with post-capitalist society we want to 
discover the changed patterns and conditions of class formation and 
class conflict. Historically, the problem of an analysis of post-capitalist 
society in terms of class may be formulated as one of the destiny of the 
"old" conflict between capital and labor, bourgeoisie and proletariat. 
If we project the historical problem into the present it becomes trans- 
formed into the task to apply the tool of class theory to some critical 

248 Classes in Post-Cafitalist Society (I) 

features of post-capitalist society and to try to contribute in this way 
to the understanding of the society in which we live. 

The "society in which we live" covers a multitude of generalities. 
It is as awkward as it can be fruitful to lump together, in sociological 
analysis, a number of societies under a general term, such as "advanced 
industrial" or "post-capitalist society." Most of the data presented 
so far in this study relate to contemporary British, American, and Ger- 
man society. It is an open question whether these data, or the con- 
clusions derived from them, apply to French, Italian, Japanese, or 
Russian society as well — and indeed, whether there are not significant 
differences between Germany, Britain, and the United States which 
would have to be taken into account. I am well aware of this problem, 
and of the criticism to which I lay myself open in not discussing it 
more elaborately. It is nevertheless my intention to try to consider 
some of the features of the industrial and political life of "post-cap- 
italist society" without referring to specific countries or periods in 
more definite terms than by stating my belief that the conclusions of 
our analysis apply at least to those democratic countries of the West 
that underwent industrialization in the nineteenth century, and at 
most to all societies at an advanced stage of industrial development. 
In this analysis, I shall avoid generalities by specifying subject rather 
than time and place. By concentrating on a few salient points, I hope 
to pave the way for more detailed investigations. This essay — for such 
it is — does not pretend to answer all problems of conflict in post-capi- 
talist society. 


One social institution to which Marx devoted a great deal of at- 
tention has survived capitalist society: the industrial enterprise. Tri- 
vial as this statement may sound, there is no reason to avoid it. There 
can be no doubt that many changes have occurred in the century be- 
tween 1 850 and 1950 both outside and inside the industrial enterprise. 
It may appear meaningless to identify the small factory of a capitalist 
entrepreneur in 1850 with the large corporation of 19 50 in terms of 
productive capacity and number of employees, technical perfection 
and spatial extension, complexity of organization and conditions of 
work. However, although these changes are by no means irrelevant 
for conflict analysis, we have to start with a more fundamental rela- 
tion which remains, or has remained so far, unchanged. In capitalist 
as in post-capitalist society, in the Soviet Union as in the United States, 

classes in Post-Caf kalis t Society (I) 249 

the industrial enterprise is an imperatively coordinated association.^ 
Everywhere it displays those conditions of social structure which give 
rise to social conflict in terms of class theory. Wherever there are 
industrial enterprises, there are authority relations, latent interests, 
quasi-groups, and (Industrial) classes. 

In dealing with the formal organization of the enterprise a dis- 
tinction is usually made between the "functional" aspect of the division 
of labor and the "scalar" aspect of super- and subordination. Both are 
functionally necessary 5 they are complementary aspects of industrial 
organization. One of the secrets of the increase of productivity by 
mechanized factory production lies in the subdivision of the total 
process of production into cooperative detail processes. Every one of 
these is equally indispensable for the accomplishment of the total 
process. From a strict functional point of view, the unskilled laborer, 
the foreman, and the executive stand on one level ; the enterprise can- 
not function if one of these positions remains vacant. However, for 
purposes of the organization, coordination, and leadership of such 
subdivided detail processes a principle other than the division of labor 
is needed. A system of super-and subordination guarantees the fric- 
tionless operation of the total process of production — a system, in 
other words, which establishes authority relations between the various 
positions. The incumbents of certain positions are endowed with the 
right to make decisions as to who does what, when, and how 5 the in- 
cumbents of other positions have to submit to these decisions. Nor 
are the commands given and obeyed in the industrial enterprise con- 
fined to technical work tasks : hiring and firing, the fixing of wage rates 
and piecework systems, introduction and control of disciplinary regu- 
lations, and other modes of behavior are part of the role expectations 
of the incumbents of authority positions in the enterprise and give 
rise, therefore, to its scalar or authority structure. For the industrial 
worker, the labor contract implies acceptance of a role which is, inter 
alia, defined by the obligation to comply with the commands of given 
persons. Industrial authority does not, of course, involve the subordi- 
nation of total persons under other persons j it is restricted to persons 

^ I deliberately neglect in the following analysis the often overstressed tendencies 
of automation and their social implications. It is of course conceivable that automa- 
tion will change not only the conditions of work, but also the authority structure of the 
enterprise, but this is today without doubt a distant dream which need not and must 
not be taken into account, unless we want sociological analysis to evaporate into the 
skies of fantasy or science fiction. 

250 classes in Post-Capitalisf Society (I) 

as incumbents of given, limited roles j but it is therefore no less author- 
ity, i.e., a "probability that a command with a given specific content 
will be obeyed by a given group of persons." Although, in other 
words, the foreman cannot legitimately command his workers to col- 
lect stamps in their leisure time, there exist in the industrial enterprise, 
within a definable range, authority relations in the strict sense of class 

Some industrial sociologists have denied the applicability of the 
category of authority to industrial enterprises either in general or in 
specific instances. Thus, Neuloh confronts "unilateral industrial or- 
ganization" characterized by authority relations with the (sociologi- 
cally contradictory) notion of "bilateral organization," where "super- 
ordinates and subordinates or their representatives participate with 
equal rights in the process of decision-making" (160, p. 54), so that 
the "subordinates" become "superordinates" and there are no real 
"subordinates" at all, Mueller goes even further if he states cate- 
gorically that "relations of subordination in the enterprise" are es- 
sentially "not authority relations" (158, p. 171). "The thesis . . . 
that the enterprise is a ^sphere of authority' must ... be rejected," 
Mueller asserts (p. 172), although a little later he remarks that 
"super- and subordination in industrial organization is essential to the 
enterprise" (p. 173). This latter remark shows quite clearly that 
Mueller is concerned not with the fact but with the concept of author- 
ity. The same holds for Neuloh, who states with excessive caution at 
the end of his work: "It would seem to be a general conviction that 
total coordination [i.e., total lack of subordination — R.D.] is impos- 
sible in any kind of enterprise, particularly in the industrial enter- 
prise," since this would "involve the abolition of hierarchies in the 
enterprise, i.e., of any discipline and authority" ( 1 60, p. 249) . This is, 
indeed, not merely general conviction, but an indispensable condition 
of the structure of the enterprise. By contrast, it is "general convic- 
tion" that Mueller's "super- and subordination" and Neuloh's "dis- 
cipline and authority" describe authority relations in the strict sense 
of the term, and that there is no reason not to call them by this name. 
Schelsky rightly emphasizes the "particular authority foundation" of 
the industrial enterprise as distinct from such structures as are founded 
on technical requirements (164, p. 87). "Wherever enterprises are 
set up, a few command and many obey," Bendix states as a matter of 
course in the second sentence of his work on industrial authority (138, 
p. i). Later he repeats more specifically: "All economic enterprises 

Classes in Post-Capitalist Society (I) 2 Si 

have in common a basic social relation between the employers who 
exercise authority and the workers who obey" (p. 13). However one 
may resent the word authority, substantively there can be no doubt 
that industrial enterprises are at all times and in all places imperatively 
coordinated associations. In this sense Mueller is involuntarily right 
when he refers to Marx and remarks : "Even Karl Marx had to admit : 
*A11 immediate social or communal labor on a larger scale requires 
more or less leadership. . . . An individual violinist conducts him- 
self, but an orchestra requires a conductor' " (158, p. 170). 

Since the industrial enterprise has an authority structure and is 
therefore an imperatively coordinated association, we are entitled to 
assume that the incumbents of positions of domination and subjection 
within it are united in two conflicting quasi-groups with certain latent 
interests. This inference follows from the model of class formation. 
If the theory of group conflict proves useful, its validity is as universal 
as the imperative character of the enterprise itself. Wherever there 
are industrial enterprises, there is a quasi-group of the incumbents of 
roles of domination, the latent interests of which are in conflict with 
those of a corresponding quasi-group of incumbents of roles of sub- 
jection. If W. E. Moore states that "the interests of management and 
labor may be 'basically the same,' and yet the relations between the two 
groups be anything but harmonious" ( 1 57, p. 400) and adds that "in 
a complete absence of like interests, there can be no conflict: there is 
nothing to fight about" (p. 399), he merely expresses the fact of the 
two faces of society with specific reference to the enterprise. Like all 
other units of social structure, industrial enterprises may be described 
in terms of integration theory. Their stability presents itself, then, as 
based on a "common value system" or "common interests." But for 
certain purposes of analysis this approach is insufficient j we need the 
parallel approach of coercion theory. From the point of view of 
coercion theory, however, the interests of the incumbents of different 
authority positions appear as conflicting. Thus, Taylor's "firm con- 
viction that the true interests of the two [management and labor] 
are one and the same" (167, p. 10), that employers and workers are 
not opponents but partners, does not constitute an objection to class 
analysis. As opposed to Taylor, we do not postulate any "true in- 
terests." We merely assert that Taylor's exclusive emphasis on the 
community of interests among all participants of the enterprise is 
plainly insufficient for the explanation of certain phenomena, such as 
strikes, and that it is therefore necessary to assume a conflict of latent 

252 classes in Post-Cafitalist Society (I) 

interests in the enterprise emerging from the differential distribution 
of authority. 

Speaking of "common interests," Taylor and others usually spec- 
ify the substance of these interests: "In the ordinary course of affairs 
it is mutually advantageous to management and labor for an industrial 
establishment to continue in operation" (Moore, 157, p. 400). Anal- 
ogously, the question of the substance of conflicting latent interests in 
industry must be raised. This question may appear unnecessarily ab- 
stract at a time at which everybody knows that the conflict of interests 
between employers and workers manifests itself in the form of dis- 
putes over wages and conditions of work. Yet it is questionable 
whether the substantive interest in higher wages or higher profits 
defines the substance of the latent interests of industrial classes suffi- 
ciently clearly. In an acute argument Drucker has tried to reduce 
wage disputes to their "real basis": "The wage rate is the traditional 
symbol for the real conflict rather than the issue itself. The basic prob- 
lem is a conflict between the enterprise's view of wage as cost, and the 
employee's view of wage as income. The real issue is not, properly 
speaking, an economic one, but one over the nature and function of 
wage: shall the need of the enterprise or the need of the employee 
be the basis for determining the function of wage?" (144, p. 58). 
Drucker comes close to our problem, but not quite close enough. In 
reducing known manifest interests to latent interests at their basis we 
must abstract even further from substantive contents connected with 
specific social conditions. "It is clear that a conflict is always con- 
cerned with a distribution of power. An exertion of power is necessary 
in order to retain a share in the determination of future relations, as 
well as for the acquisition or retention of other benefits which may be 
the immediate ^reasons' for the conflict. This is to say that the im- 
mediate and necessary goal of any conflict is complete or partial vic- 
tory" (Moore, 157, p. 400). If we formulate, as we must, the latent 
interests of industrial quasi-groups on an equally formal and general 
level as the "common interest" in the continued operation of the en- 
terprise, their substance can be described as the maintenance or change 
of the status quo by conservation or modification of existing relations 
of authority. 

In disputes between trade unions and employers, an argument is 
often put forward by the employers which has found its way into 
sociological literature also. Employers like to assert that they repre- 
sent the interests of the total enterprise whereas the unions merely 
stand for partial interests. It might appear that there is no convincing 

Classes in Post-Cafitalist Society (I) 253 

refutation of this argument.* However, in the light of the theory of 
group conflict it becomes apparent what the basis of this argument is 
and why it is ideological, i.e., demonstrably false — a fact which is of 
considerable significance for class analysis. We have seen earlier that 
the interests of a ruling group assume, as ruling interests, the char- 
acter of accepted values in a unit of social structure. They are a re- 
flection of the real structure, the existing conditions, although these 
are upheld and guaranteed by the rule of but one class. They might 
therefore appear as binding for all elements of a unit of social struc- 
ture. Yet the theory of class exposes the fact that the existing condi- 
tions are themselves in a sense merely "partial," that they exist by 
virtue of the authority of one part, or class. As the prime minister is 
both representative of the whole nation and exponent of the majority 
party, the entrepreneur is both "the enterprise" and one partial in- 
terest in the conflicts generated by its structure — depending on the 
image of society underlying our analysis. In this sense, "conservation" 
and "modification" of a status quo are, from the coercion point of 
view, strictly equivalent "partial" interests the conflict of which can 
be conceived as one of the determinants of the dynamics of social 

Two objections are frequently raised these days against the uni- 
versal reality of conflicting latent interests in industry. They are easily 
disposed of, yet it may further clarify the issue to discuss them briefly. 
The first of these objections is based on the thesis that what is often 
called the "bourgeoisification of the proletariat," i.e., the improve- 
ment of the economic situation of industrial workers, makes the as- 
sumption of continuing conflict unreasonable and, indeed, nonsensical. 
If, it is argued, the workers are no longer proletarians, if they do not 
live in poverty and suppression, they no longer have reason to revolt 
against their employers. The public of post-capitalist societies realizes, 
with baffled surprise, the continued reality of strikes and yet insists, at 
the same time, on the theory that industrial conflict has lost its causes 
and issues where the standard of living is high. This paradox testifies 
to a remarkable consistency of conviction, if not to insight. The theory 
of group conflict does not postulate any connection between class con- 
flict and economic conditions. For the emergence of social conflicts the 
standard of living of their participants is in principle irrelevant, for 
conflicts are ultimately generated by relations of authority, i.e., by the 

* At any rate, many trade union leaders have found it difficult to argue against 
this point, except by dogmatic insistence on their power. Cf. 42. 

254 Classes in Post-Cafitalist Society (I) 

differentiation of dominating and subjected groups. Even if every 
worker owns a car, a house, and whatever other comforts of civilization 
there are, the root of industrial class conflict is not only not eliminated, 
but hardly touched. The fact that economic demands may provide 
the substance (a substance situationally specific and in that sense in- 
cidental) of manifest interests must not give rise to the erroneous 
notion that satisfaction of these demands eliminates the causes of con- 
flict. Social conflict is as universal as the relations of authority and im- 
peratively coordinated associations, for it is the distribution of author- 
ity that provides the basis and cause of its occurrence. 

A second objection against the assumption of the persistence of a 
latent conflict of interest consists in the thesis that the replacement of 
capitalists by managers has removed the basis of industrial class con- 
flict. Upon closer inspection, this thesis, too, proves untenable. As we 
have seen, latent interests can be conceived of as quasi-objective role 
expectations. They are held not by persons, but by positions, or by 
persons only insofar as they occupy certain positions. If a person oc- 
cupies a position of domination in an enterprise, it is irrelevant in 
principle^ whether his authority is based on property, election by a 
board of directors, or appointment by a government agency. For the 
latent interests of the incumbents of positions of authority, their in- 
cumbency of these positions is the sole significant factor. Although, 
therefore, their modes of recruitment and bases of legitimacy make 
for significant differences between capitalist and manager in other con- 
texts, their authority positions in the enterprise are alike, and their 
places in conflicts of interest identical. For the explanation of group 
conflict, the factual relations of authority are the crucial factor. To this 
extent I agree with Burnham's thesis and with Marx's and Renner's 
analysis of joint-stock companies j the replacement of functioning 
owners or capitalists by propertyless functionaries or managers does 
not abolish class conflict, but merely changes its empirical patterns. In- 
dependent of the particular personnel of positions of authority, in- 
dustrial enterprises remain imperatively coordinated associations the 
structures of which generate quasi-groups and conflicting latent in- 

Apart from such inconsequential objections, there is one rather 
difficult problem, which becomes apparent if we try to define the bor- 

^ If I refer here and elsewhere to the "irrelevance" of certain social changes, I 
always mean irrelevance in terms of the theoretical model of class analysis. This is of 
course not to say that the changes in question may not be relevant as intervening 
variables affecting the intensity or violence of conflicts. 

Classes in Post-Capiialist Society (I) 255 

derline between the quasi-groups of industrial associations. Bendix 
remarks that in every enterprise there are "a few" who command and 
"many" who obey. Presumably, he has the capitalist enterprise in 
mind when he says so. But the same criticism applies to his statement 
as was raised against Pareto's and Mosca's thesis of a ruling "mi- 
nority." In the enterprises of post-capitalist society, authority is typi- 
cally no longer exercised by one individual, or even by a few. A com- 
plex system of delegation of responsibility obliterates virtually to 
invisibility the dividing line between positions of domination and sub- 
jection. Even if we ignore for the moment the place of labor repre- 
sentatives in the formal organization of the enterprise, there are two 
groups that stubbornly resist allocation to one or the other quasi-group. 
One of these consists of the so-called "staff" of the enterprise, the en- 
gineers, chemists, physicists, lawyers, psychologists, and other special- 
ists whose services have become an indispensable part of production in 
modern firms. Occasionally, these specialists have a defined place in 
the scalar structure of authority in the enterprise, usually by the in- 
clusion of specialist qualifications in the definition of the expectations 
associated with managerial roles." More frequently, however, the 
staff is linked with the line of authority by an intricate system of cross- 
relations without its members having immediate authority except over 
their secretaries and assistants ("line-staff system"). In this case the 
class situation of specialists in the enterprise remains as uncertain as 
the class situation of intellectuals in society. They are neither super- 
ordinates nor subordinates j their positions seem to stand beyond the 
authority structure. Only insofar as they can be identified as (often 
indirect) helpers of management, can they be called a marginal part 
of the ruling class of the enterprise. 

The number of staff positions in industrial enterprises is usually 
fairly limited. But there is another troublesome group of positions in 
the modern enterprise which is far more numerous, namely that of all 
salaried employees, of department heads and typists, accountants, and 
foremen, etc. Few of these positions can be described as positions of 
subjection in the sense in which those of workers are. But are they 
positions of domination? Do salaried employees participate in the 
exercise of authority? Do they belong to the ruling class of the in- 
dustrial enterprise? In the first place, the distinction introduced 
earlier between bureaucrats and white-collar workers is here of im- 

" Not infrequently, it is expected of the technical manager today that he be a 
trained engineer, of the business manager that he be a trained economist, of the gen- 
eral manager that he be a lawyer, etc. 

256 classes in Post-Capiialist Society (I) 

portance. The latter clearly belong to the quasi-group of the subjected 
and do not therefore present a special problem. But what about the 
bureaucrats? In the following chapter we shall have to investigate 
the place of bureaucrats in social conflict rather more extensively. An- 
ticipating the result of this investigation, however, I think that with 
certain qualifications the conclusion is indicated and, indeed, forced 
upon us that the positions of bureaucrats in the enterprise have to be 
counted among the positions of domination. The theory advanced by 
Renner and others since the turn of the century, and called "theory of 
delegation" by Croner, provides the only consistent explanation of the 
authority position (although not of the social status or economic situa- 
tion! ) of bureaucratic roles. Bureaucratic roles in the enterprise must 
be described as difFerentiated management roles. The total process 
of the exercise of authority appears, in the modern enterprise, sub- 
divided into a multitude of positions coordinated by a peculiar type of 
organization. In the analysis of bureaucratic roles below we shall see 
that this subdivision and coordination bears striking similarities to the 
division of labor in technical production. In both cases, differentiation 
has the efFect of splitting up a totality into a multitude of elements to 
the extent of almost entirely alienating each of the elements from the 
totality of which it is a part. Here, appearances are deceptive, for the 
fact remains that industrial bureaucracy or, as Renner calls this group, 
the "service class," stands as a whole on this side of the borderline 
which separates the possessors of authority in the enterprise from the 
subjected workers, both manual and white-collar. By virtue of their 
positions, bureaucrats are members of the ruling class of industry and 
share its latent interests.^ 

It may appear superfluous to devote so extensive a discussion to 
the structural origins of industrial conflict in view of the fact that the 
organized interest groups of trade unions and employers' associations 
belong to the accepted institutions of post-capitalist society. Yet this 
appearance is deceptive. Only if we manage to identify the structural 
origins of industrial conflict unambiguously can we hope to be able to 
assess the place of organized interest groups in post-capitalist society 
with sufiicient accuracy. The authority structure of the industrial en- 
terprise generates, in every social order and at all times, independent 
of the socio-economic status of labor and the modes of recruitment of 

^ Clearly, there are nevertheless significant differences, even from the point of 
view of social conflict, between managers and bureaucrats. How these affect analysis 
in terms of conflict theory will be shown in the discussion of bureaucratic roles in 
Chapter VIII. 

classes in Post-Capitalist Society (I) 257 

management, latent interests and quasi-groups in the sense of class 
theory. Their conflict may be described as the ultimate source of class 
antagonisms in industry. At the same time, this statement leaves out 
of consideration whether industrial conflict becomes acute or remains 
latent, whether it assumes regulated patterns or is of the civil-war 
type, whether it is confined to industry or molds the whole of society. 
Empirically, these are the crucial questions, and the answers to them 
cannot be derived from the authority structure of the enterprise. For 
them to be satisfactorily answered, we now have to look at the em- 
pirical conditions of industrial conflict in post-capitalist society. 


In connection with the brief re-analysis of Marx's data above, I 
have indicated that there is a tendency toward decreasing intensity and 
violence in industrial conflict by virtue of the "institutionalization of 
class conflict" and the development of "industrial democracy." This 
assertion must now be made specific and documented. Without en- 
tering into a detailed historical account of the development of con- 
flict patterns in various countries, I shall try in this section to trace some 
of the typical (and universal) elements of conflict regulation in in- 
dustry as they aflFect the violence of manifestations of conflict. My 
thesis is that in post-capitalist society industrial conflict has become less 
violent because its existence has been accepted and its manifestations 
have been socially regulated. Today, industrial conflict is recognized 
as a necessary feature of industrial life. This recognition, as well as 
the establishment of regulatory institutions, constitutes in itself a 
structure change which is due to no small extent to the efi^ects of in- 
dustrial conflict. 

The establishment of "industrial democracy" consists of a number 
of structural arrangements, each of which warrants brief inspection. 
In the following pages, I propose to analyze five elements of industrial 
democracy in advanced countries which seem to me of particular im- 
portance: (i) the organization of conflicting interest groups itself j 
(2) the establishment of "parliamentary" negotiating bodies in which 
these groups meet j (3) the institutions of mediation and arbitration j 
(4) formal representations of labor within the individual enterprise j 
and (5) tendencies towards an institutionalization of workers' partici- 
pation in industrial management. The purpose of the following anal- 
ysis, as of the whole chapter, is to stimulate thought and research 
rather than to present conclusive solutions j thus, we shall concentrate 
on significant facts rather than present all relevant data. 

258 classes in Post-Capitalist Society (I) 

( I ) In the sphere of industry we encounter the conflicting classes 
in full formation. In a way, the conditions of industry in post-capital- 
ist society might be described as an empirical analogue of the model of 
class theory, an ideal type come true. The authority structure of the 
enterprise generates the two quasi-groups of management and labor, 
along with their latent interests j from these are recruited the interest 
groups of employers' associations and trade unions, with their specific 
manifest interests. For many decades, now, the disputes between trade 
unions and employers' associations have presented the well-known 
picture of industrial conflict. 

There are a number of problems in connection with the formation 
of industrial classes which, although well worth investigating, we can 
only touch upon here, (a) It would be interesting to investigate why 
trade unions typically attempt to recruit all members of the quasi- 
group of labor into their organization ("closed shop") and to actual- 
ize their interests permanently, whereas employers' associations are 
much looser organizations composed of a few representatives of their 
underlying quasi-group. One would have to inquire whether this 
pattern indicates a general regularity of manifest conflict, or whether 
it might not be conceivable that a subjected class is also represented by 
a small interest group of delegates.^ (h) Complex problems are raised 
by the relation between employers' associations and industrial bureau- 
crats, particularly their union-type organizations, if we introduce the 
premise that the service class is part of the ruling class, (c) In many 
countries, the relation between different (often conflicting) interest 
groups based on the same quasi-group, such as Christian and Socialist 
trade unions, presents important problems, (d) Finally, we cannot 
here discuss the extremely significant consequences of the fact that 
organized trade unions are themselves imperatively coordinated as- 
sociations which potentially generate within themselves the same con- 
flicts between dominant and subjected groups ("bosses" and members) 
as other associations. 

Other problems might be added to these. However, we are here 
concerned merely with a more limited question: in what sense does 
even the organization of conflict groups itself contribute to a decrease 
in the intensity and violence of class conflicts? It is worth pursuing the 

* There is some indication today that trade unions develop from membership 
organizations to offices staffed with a few functionaries who professionally represent 
the interests of labor. It is just possible that the trade union of the future will look 
more like American than like European political parties, i.e., that it will be a "latent" 
organization actuated only for special purposes. 

classes in Post-Capitalist Society (I) 259 

subject that every act of organization is as such a process of institu- 
tionalization. The organization of an interest group creates an entity 
in time, endowed with a "charter," certain norms, a material instru- 
mentarium, and a personnel that constitute — if we accept Malinow- 
ski's definition — an institution. This fact has certain prerequisites j 
and certain consequences flow from it. Thus, the organization of trade 
unions presupposes, apart from the presence of a quasi-group of latent 
interests, those technical, political, and social conditions which we have 
described as conditions of organization and the presence of which sig- 
nifies in itself a minimum of recognition of the legitimacy of conflict 
and its issues. The organization of trade unions is in this sense the 
first structure change of industrial society caused by class conflict j by 
its completion class conflict loses some of its intensity and violence. 

This conclusion is supported by the consequences of the organiza- 
tion of interest groups. Organized groups stand in open, and there- 
fore in controllable, conflict. As long as the organizations of sub- 
jected groups are not dissolved, their political prerequisites not abol- 
ished, absolute suppression of this class is no longer possible. Or- 
ganized interest groups of subjected classes have means to enforce 
recognition of their interests at their disposal j moreover, they are in 
principle accessible for negotiations, i.e., regulated disputes. One 
cannot negotiate with unorganized, loosely connected "rebels" 3 for 
conciliation comprehensive organization growing out of a quasi-group 
is indispensable. Thus, the conclusion is suggested that the "democ- 
ratization" of industrial conflict begins with the organization of in- 
dustrial classes. 

(2) In fact the formation of industrial interest groups has soon 
been followed, in all industrial countries, by the emergence of ne- 
gotiating bodies in which the representatives of both parties convene 
for settling their disputes by discussion. In some countries these ne- 
gotiating bodies, originally formed ad hoc whenever necessary, have 
stabilized into persistent institutions j there is a machinery for collec- 
tive bargaining and joint negotiation. "The term ^collective bar- 
gaining' is applied to those arrangements under which the wages and 
conditions of employment are settled by a bargain, in the form of an 
agreement made between employers or associations of employers and 
workpeople's organizations" (156, p. 15). The structure of these 
negotiating bodies differs in different countries. "The area regulated 
by collective contract is, for example, much wider in America than in 
Germany. In the latter country many of the relations between em- 
ployer and employee are regulated on the basis of mutual under- 

26o classes in Post-Cafitalist Society (I) 

standing without this anywhere being put down in writing j many 
other conditions are fixed preferably by law and less by collective con- 
tracts" (McPherson, 155, p. 69). Apart from this latter qualifica- 
tion, the significance of collective bargaining for industrial conflict 
can be asserted in general for all post-capitalist societies. 

Occasional or statutory negotiating bodies of employers' associa- 
tions and trade unions serve a quasi-parliamentary function in indus- 
trial class conflict. In them the representatives of both parties meet 
in order to articulate their disputes on the basis of certain rules of the 
game (such as an order of procedure, but also according to other norms 
and forms of behavior)® and, if possible, to arrive at a common de- 
cision. T. H. Marshall has demonstrated convincingly how radically 
this structural principle differs from a private-law conception of the 
labor contract (cf. 57). For us, the essential fact is that by collective 
bargaining the frozen fronts of industrial conflict are thawed. If the 
representatives of management and labor meet regularly for ne- 
gotiations, gradual changes of social structure replace the tendency 
toward revolutionary explosions and civil war. 

(3) However, this possibility is by no means entirely removed 
where collective bargaining bodies exist. The rules of the game of 
political democracy can only partly be applied to the industrial con- 
flict of interests. "The whole of this collective system rests upon the 
principle of mutual consent, and the value of the agreements and the 
machinery for settling disputes has depended upon the loyal accept- 
ance by the constituent members on both sides of the decisions reached" 
(156, p. 16). Not only the acceptance of agreements once reached, 
but even more the difficulty of reaching agreement endangers the ef- 
fectiveness of conciliation. Majority decisions are structurally im- 
possible in the negotiating bodies of employers' associations and 
unions; decisions are reached unanimously or not at all. To the extent 
to which agreement between conflict groups cannot be attained by ne- 
gotiation, industrial conflict threatens the democratic process itself. 
Violent conflict — strike and lockout — remains its background and, 
often, its result. 

In order to remove the possibility of outbreak of violent conflict 
even further, most advanced industrial societies have developed a 

^ These are nicely brought out when a bitter wage dispute which has no result 
ends in an exchange such as that found in the minutes of an English negotiation which 
1 have analyzed elsewhere (42), where the union representative (a man known for 
his radical views) says, "I presume we say Good morning then!" and the employers' 
representative replies, "I am sorry, Mr. T." 

Classes in Post-Cafitalist Society (I) 261 

second line of institutional safeguards, a system of mediation and 
arbitration. It is the declared aim of arbitration to make a last non- 
violent attempt at regulating conflicts in case of breakdown of con- 
ciliation. In connection with the general discussion of conflict regu- 
lation I have presented a schema of types of third-party interference 
in conflict, ranging from (voluntary) mediation through (partly com- 
pulsory) arbitration to compulsory arbitration. Apart from the last 
type, such interference may operate as an effective means of reducing 
the violence of industrial conflict. There are many empirical illustra- 
tions of this fact. However, for arbitration to be effective, it is neces- 
sary that it be based on the acceptance of the conflict itself, i.e., that 
it follow a political rather than a legal conception of its task, to use 
Lockwood's distinction (cf. 86). If, and only if, mediation and arbi- 
tration are conceived of as mechanisms for facilitating legitimate de- 
cision-making on the part of both parties to a conflict, do they con- 
tribute to industrial democracy. 

(4) The three factors mentioned so far form a coherent pattern. 
Its rationale lies in the autonomous, in that sense democratic, regula- 
tion of conflict. This pattern can be found in most post-capitalist so- 
cieties, but particularly in Britain and the United States. The two 
remaining factors, representation within the enterprise and co-deter- 
mination, also form a coherent pattern. Again, traces of the pattern 
are present in most industrial countries, but it is particularly promi- 
nent in Germany. Its most general principle can be described as the 
attempt to institutionalize industrial conflict by modifying the author- 
ity structure of the enterprise itself. So far as we can see today, the 
range of variability of possible modifications of the authority structure 
of the industrial enterprise is relatively limited.^" Yet there is the 
possibility of supplementing the authority structure by what Schelsky 
called an "institutionalized side-hierarchy of employees' representa- 
tives" (165, p. 187), such as shop stewards, comites d^entreprisey 
Be trieb sr ate y or, in Communist countries, party cells in the enterprise. 
Schelsky states on the basis of this important structure change of mod- 
ern industry that "a dualistic authority structure belongs to the hierar- 
chical constitution of the modern enterprise" (165, p. 185). It is 
worth following up some of the consequences of this development for 
violence and intensity of industrial conflict. 

In Germany, labor representation within the enterprise found its 

^° This conclusion is brought out with great clarity by Bendix in his study of in- 
dustrial authority (138). 

262 Classes in Post-Capitalist Society (I) 

way into legislation as early as 1920. Special legislation defines the 
task of shop councils" by two main expectations. Shop councils are 
established "in order to defend the common economic interests of 
employees . . . vis-a-vis the employer, and in order to support the 
employer in realizing the functions of the enterprise." The striking 
ambiguity of the position of shop councils — and, to a lesser extent, of 
British and American shop stewards — becomes apparent in this formu- 
lation. The first part of the legal definition, ascribing to shop councils 
the task of defending the interests of labor, seems to bear out Schel- 
sky's and Drucker's comparison of this institution with a parliamentary 
opposition. From this point of view, the institution of shop councils 
neither modifies nor supplements the authority structure of the en- 
terprise j leaders of opposition have no legitimate power within the 
state. There is, however, a second aspect of the position of shop coun- 
cils. Apart from defending the interests of labor, shop councilors 
have also to give certain commands to the workers j they have to "sup- 
port the employer in realizing the functions of the enterprise." More 
apparently than with the shop stewards of English and American in- 
dustry, German shop councils occupy a curiously ambiguous place in 
the enterprise. They are both opposition and side-government, rep- 
resentatives of the interests of labor and agents of management. For 
this reason, much the same problems are associated with their roles as 
Roethlisberger and others found characteristic for the foreman. At 
least in Germany, the shop councilors are, even more than the fore- 
man, "men in the middle," "masters and victims of double talk."^^ 
In so far as the representation of labor within the enterprise fol- 
lows the double definition of German shop councils, its effect on the 
intensity and violence of industrial conflict is as ambivalent as its posi- 
tion in the authority structure of the enterprise. On the one hand, class 
theory would lead one to assume that the establishment of institutions 
designed to "defend the common economic interests of employees" 
serves to reduce the violence of conflict by providing peaceful chan- 
nels of expression for antagonistic interests. On the other hand, the 

^^ The German term Betriebsrat will in the following analysis be translated by 
"shop council," so as to avoid confusion with the shop steward system, which is dif- 
ferent in several respects. The following quotation stems from the Betriebsrätegesetz 
of 1920, section i of which defines the functions of shop councils. 

^^ This situation is most pronounced in German industry. However, the com- 
parative analyses by McPherson (155) and A. Philip (161, Chapter II) show that 
shop stewards and comites i* entrefrise also serve certain managerial functions, so that 
the present analysis applies to other countries as well. 

classes in Post-Cafitalist Society (I) 263 

very design of shop councils tends to alienate this institution from 
those whose interests it is supposed to serve. By necessity, the repre- 
sentatives of labor get entangled with the tasks, decisions, and — at 
least indirectly — interests of management. What appears to be an 
interest group of the subjected turns out to be an agent of the domi- 
nating group as well. It seems probable that this kind of perverted 
conflict regulation will increase rather than diminish both the violence 
and the intensity of conflict by simultaneously opening and blocking 
one of its channels of expression. 

This latter point is most clearly illustrated by the relationship of 
labor representatives within the individual enterprise to the larger or- 
ganizations of labor, i.e., the trade unions. In England and the United 
States "shop stewards are appointed by and subject to the control of 
their unions" (156, p. 66). They are in this sense delegates of the 
comprehensive interest group of labor within the enterprise. As such, 
they are symptom and tool of the institutionalization of class conflict 
and serve to decrease its violence. Although in actual fact there are in 
Germany, also, many enterprises where the trade unions have at least 
informal control over selection, election, and work of the shop coun- 
cils, these are in principle independent of the unions. There are there- 
fore at least two distinct interest groups of labor,^^ the relations be- 
tween which may be friendly, competitive, or conflicting. The theory 
of group conflict would suggest that the closer the connection between 
labor representatives in the enterprise and comprehensive interest 
groups, the better do these interest groups serve their purpose of 
regulating industrial conflict and thereby reducing its violence. The 
shop councilor, as an incumbent of roles of domination, becomes part 
of the ruling class of industry, deprives labor of one of its channels 
of expression, and provides the cause for new conflicts of the class type 
within the enterprise and industry as a whole. 

(5) This conclusion holds a fortiori for a final element of indus- 
trial democracy which we have to discuss and which, once again, is most 
prominent in West German industry: workers' participation in man- 
agement. Apart from strictly syndicalist tendencies in a number of 

^^ Strictly speaking, the shop council cannot be called an interest group but is 
merely the head of an otherwise nonexistent group. What happens is that the compre- 
hensive quasi-group of labor functions as such with respect to trade unions, and is at 
the same time split up into innumerable smaller units and recruiting fields within the 
walls of individual enterprises. Thus, every wori<er functions as industrial worker and 
as worker of factory X, or as (potential) trade union member and as (potential) shop 

264 classes in Post-CafkaUst Society (I) 

historical and contemporary societies, the coal, iron, and steel industry 
of post-war Germany provides the first large-scale example of an 
attempt to furnish labor with a share in the management of indus- 
trial enterprises. There have been numerous studies, sociological and 
otherwise, of the German co-determination experiment in recent 
years, and I do not intend to review here these investigations or the 
problems they have raised. There is, however, one element of co- 
determination which is of immediate relevance for our problem and 
requires some consideration. One of the crucial stipulations of the 
"Law Concerning Co-Determination of Employees in the Boards of 
Directors and of Executives in the Enterprises of Coal Mining and 
the Iron and Steel Industry of May 21st, 1951" relates to the in- 
clusion of a labor representative in the executive board. Section 13 
of the third part of the law, dealing with executive boards, says: 
"i. A labor manager {Arbeitsdirektor) is appointed with rights equal 
to those of the other members of the legally representative executive 
board. The labor manager cannot be appointed against the votes of 
those members of the board of directors elected in accordance with 
section 6 [i.e., the representatives of labor in the board of directors 
— R.D.]. ... 2. The labor manager, like the other members of 
the legally representative executive board, has to fulfill his tasks in 
close cooperation with the total board." What are the structural con- 
sequences for industrial conflict of this stipulation, and the real con- 
ditions created on its basis? 

Two conclusions can immediately be derived from the wording 
of the law. First, this law stipulates the creation of a new position. 
It does not define in any detail the role, the behavior expectations 
attached to this position. However, its designation "labor manager" 
implies an indication that the legislators had in mind an executive 
who is primarily concerned with personnel matters in the widest 
sense. ^* The law states quite explicitly, however, that the role of the 
labor manager is a role of domination, a leadership or management 
role equipped "with equal rights." The position of labor manager 
is clearly defined as belonging to the ensemble of positions endowed 
with the right to issue authoritative decisions and commands. Sec- 
ondly, the wording of the law stipulates a mode of recruitment of 
persons to this position. Here, again, the text remains purely formal. 
It is not suggested, e.g., that the labor manager has to be a (former) 

■^* To the present day, the German trade unions eagerly disclaim this implica- 
tion, which, in their opinion, puts the labor manager at a disadvantage with his col- 
leagues. Despite this fact, I should maintain my interpretation. 

Classes in Post-Capitalist Society (I) 265 

worker or trade union member, although this idea may have played 
a part. It is merely said that he must have the confidence of the labor 
representatives on the board of directors, that he cannot be appointed 
against their votes. What, then, is the class position of the labor 
manager in terms of conflict theory? 

With respect to his position and role, and quite apart from its 
specific incumbent, the answer to this question is evident. As a posi- 
tion of domination it belongs to the plus-side of the zero-sum dis- 
tribution of authority positions in the enterprise j the position of labor 
manager is, to all intents and purposes, a managerial or entrepre- 
neurial position. "Like the other members of the legally representa- 
tive executive board," the labor manager belongs, by virtue of his 
position, to the ruling quasi-group of those whose objective role 
interests aim at the maintenance of existing conditions. In this sense, 
the institution of labor manager is, as Pirker justly observes, "nothing 
but the continuation of the process of rationalization of large-scale 
enterprises and of the specialization and centralization of manage- 
ment connected with it" (162, p. 417). Like his older, but similarly 
defined, colleague in English and American industry, the personnel 
manager, the labor manager is an entrepreneur. 

These facts, which are so obvious that it is truly surprising how 
rarely they are recognized, are complicated by that further stipu- 
lation of the law which prescribes a mode of recruitment. In order 
to articulate the consequences of this prescription, let us assume hypo- 
thetically the extreme case of a worker — say, a fitter — who is also an 
active unionist, being appointed labor manager of a steelworks by 
its board of directors. From the point of view of class theory, this 
means, in the first place, that a man who belongs to the quasi-group 
of the subjected becomes mobile, rises socially, and assumes a position 
that automatically puts him into the quasi-group of the rulers. His 
latent interests change from one extreme to the other. This seem- 
ingly nonsensical statement becomes reasonable if we realize that 
latent interests are associated with positions, not persons, and that 
they are therefore in principle exchangeable with positions. The in- 
dividual can abandon his class position, as he can his occupational 
position. Our fitter has become, by the decision of a board of direc- 
tors, an entrepreneur. His interests are no longer represented by a 
trade union but by an employers' association. Once again we must 
agree with Pirker, who says: "Enforcing loyalty with the union would 
be incompatible with the functions [better: the role — R.D.] of the 
labor manager which are, and must be, directed towards the optimal 

266 Classes in Post-Capitalist Society (I) 

social constitution of the enterprise within the existing economic and 
social order" (162, p. 420). 

In one respect our analysis might be charged with a certain lack 
of realism. One might ask how the fitter of our example will react 
to his new role. Will he be able or willing to forget, once he is labor 
manager, that he was a fitter before? This is a psychological question. 
However, it appears to me that we may formulate a number of al- 
ternatives of behavior that are open to the newly made labor man- 
ager of our hypothetical example, (a) He may try to continue acting 
as if he were a member of the subjected class from which he came. 
This would mean that he does not adapt himself to his new role and 
does not fulfill its tasks as he should, so that he may be a good unionist 
but will certainly be a bad labor manager, {b) On the other hand, 
he may try to adapt himself to his new role very rapidly (and succeed 
in doing so). By forgetting his origin he may indeed become an 
extreme example of an industrial ruler and soon display all the status 
symbols of a manager, {c) Finally — and this may be the most prob- 
able case — our fitter may gradually adapt himself to his new role 
as labor manager, live through what may easily be a painful conflict 
between the old and the new role, and try to solve this conflict by 
increased efforts on behalf of his former colleagues (thereby doing 
what is his task, as a labor manager, in any case). He becomes an 
entrepreneur, a member of the ruling class of industry, but he re- 
mains an employer concerned about the welfare of labor. 

I have deliberately tried to assess the significance of co-determi- 
nation in the light of class theory without reference to the unending 
discussion of the virtue or vice of this structure change. This, it seems 
to me, is the only way of gaining an impression unbiased by wishes 
or aversions. Thus, the much-discussed problem of the "conflict of 
loyalties" (union vs. enterprise) is reduced to the psychological con- 
flict of a man — who in reality is a rare exception^^ — who turns over- 

^^ No exact figures are available about the origin of the* (roughly loo) labor 
managers in West Germany. There are, however, some figures about the occupa- 
tions of labor representatives on the boards of directors (170, p. l). According to 
these, of the representatives in all co-determination enterprises in 1955, 22 per cent 
were trade union employees, 24.8 per cent higher salaried employees, managers, civil 
servants, cabinet ministers, members of parliament etc. (persons, in other words, who 
were in leading positions before), 7. i per cent experts (professors, lawyers, etc.), 14 
per cent bureaucrats, 31.5 per cent skilled workers and foremen, and 0.6 per cent 
semiskilled workers. Here, our hypothetical case would not be typical, but neither 
would it be exceptional. However, it is more than likely that the proportion of former 
workers is much smaller among the labor managers. 

classes in Post-Capitalist Society (I) 267 

night from worker to entrepreneur. Neuloh's "totally bilateral proc- 
ess of decision-making" (160, p. 58) is revealed as an extremely in- 
accurate description of the actual conditions. But it becomes evident, 
also, that the large number of workers who (according to a survey 
of the Frankfurt Institute of Social Research) expect from co-de- 
termination "equal participation in decision-making" or even "rights 
of control and veto on the part of labor" (149, p. 212), are voicing 
Utopian wishes rather than structurally justified expectations. 

In many ways, co-determination in the basic industries of West 
Germany is the most insignificant of the factors of industrial de- 
mocracy discussed in this section. This conclusion certainly holds for 
the creation of the position of labor manager. The strengthening of 
shop councils effected by a supplementary law of 1952 as well as the 
participation of labor representatives in the boards of directors stipu- 
lated by the co-determination law may be regarded as contributions 
to the effective regulation of industrial conflict. Here, negotiating 
bodies are erected or invigorated which serve to remove conflict from 
the streets to conference rooms. But the very significance of institu- 
tions of this latter type supports our general view that any attempt 
to eliminate conflict altogether is bound to fail as such and, in fact, 
intensifies existing cleavages. Regulation requires acceptance of con- 
flict j but co-determination is based on a conviction that conflict is bad 
and must be abolished. From the point of view of effective conflict 
regulation, it is an ill-conceived pattern that contradicts rather than 
supports a general trend toward the reduction of violence and in- 
tensity of industrial conflict. 


"I believe," says Parsons in one place, "that class conflict is 
endemic in our modern industrial type of society" (67, p. 333). With 
respect to industry, our analysis suggests that this belief is indeed 
well-founded. But while the existence of class conflict is indubitable, 
its manifestations have changed. The analysis of industrial democ- 
racy reveals some such changes, namely, those in the violence of inter- 
est clashes. At several points we have suggested that with decreasing 
violence the intensity of industrial conflict was also reduced. How- 
ever, the latter tendency is rather less apparent than the former. It 
is necessary, here, to define its extent and limits as precisely as pos- 
sible. In this problem, the relationship between industry and society 
is crucial. 

268 Classes in Post-Capitalist Society (I) 

We have seen that one of the reasons why industrial conflict was 
exceptionally intense in capitalist society rested with the fact that the 
lines of industrial and political conflict were superimposed. The op- 
ponents of industry — capital and labor — met again, as bourgeoisie 
and proletariat, in the political arena. Clearly, the relations between 
industry and society are close in all modern societies j in this sense, 
the term "industrial society" is fitting. At the same time, industry 
and society are, at least for purposes of analysis, discrete associations, 
the interrelations between which are not a priori definable. It is one 
of the central theses of the present analysis that in post-capitalist 
society industry and society have, by contrast to capitalist society, been 
dissociated. Increasingly, the social relations of industry, including 
industrial conflict, do not dominate the whole of society but remain 
confined in their patterns and problems to the sphere of industry. 
Industry and industrial conflict are, in post-capitalist society, institu- 
tionally isolated, i.e., confined within the borders of their proper 
realm and robbed of their influence on other spheres of society. In 
post-capitalist society, the industrial enterprise is no longer the model 
after which all other relations are fashioned. From this thesis (which, 
for this reason, one might call a "theory" by contrast to the "propo- 
sitions" or "hypotheses" derived from it) a number of significant 
consequences follov/ for class analysis. Before we consider these in 
some detail, however, the theory itself requires some elucidation. 

At an earlier point in this study we have described three sets of 
facts which would seem to suggest that Marx and Burnham are right 
in believing industry and society to be virtually identical in the modern 
world. The first of these was the purely physical extent of industry 
and industrial enterprises as well as the manifest significance of indus- 
trial production for life in contemporary society. Whatever changes 
may have occurred with respect to the relation of industry and society, 
they have evidently supported rather than contradicted this tendency. 
In ever larger enterprises an ever growing number of workers pro- 
duces ever more numerous goods. Yet this fact does not contradict 
the theory of institutional isolation of industry, as will become ap- 
parent if we look at the other two factors mentioned in our earlier 
analysis. We have stated there that the occupational role dominates 
the social position of people in industrial society, and in particular 
that of workers. Here, a counter-trend cannot be ignored. The time 
spent by the average worker on his job has decreased and will probably 
continue to decrease. Although his work plays a prominent part in 
the daily life of the worker, there are indications that other roles 

Classes in Post-Capitalist Society (I) 269 

besides his occupation gain in importance. In a sense, his occupation 
has been confined to a set place in the life of the worker, just as 
industry has been confined to a set place in the structure of society. 
We shall return to this problem presently. With respect to the sanc- 
tions available to employers to enforce conformity with the values 
of the ruling class of industry, we can note an analogous development. 
The quasi-governmental police supervision of early capitalists has 
been restricted considerably in post-capitalist society. The ruling class 
of industry rules over a rather more limited part of the lives of its 
workers J its sanctions are confined to properly industrial penalties 
and are, even here, subject to legal regulations. Employers shoot- 
ing their workers with their own police force are as hard to imagine 
today as corporal punishment, arrest, or even dismissal without rea- 
son and reduction of wages. ^"^ The loss on the part of management 
of quasi-governmental rights of control which penetrate deeply into 
the life of the individual also marks a reduction of the formerly all- 
embracing significance of industry. Here, too, confinement to a spe- 
cific, limited sphere can be demonstrated. 

That industry and industrial conflict have been institutionally 
isolated means that they have settled down in society, have found a 
stable and definite place within it. Parallels to other institutional 
spheres suggest themselves. One might argue, for example, that in 
the secularized societies of today the church has been displaced from 
the all-embracing part it played at earlier times and confined to a 
proper, defined sphere in society; the church, too, has been institu- 
tionally isolated. In all countries directly or indirectly affected by 
the French Revolution, the relation of church and state is one of de- 
limitation of competence between two separate institutional orders 
and imperatively coordinated associations. According to the thesis 
here advanced, we are witnessing, in post-capitalist society, a similar 
development with respect to industry and the enterprises of industrial 
production. Here, too, a process of confinement — that is, of delimi- 
tation of basically separate competences — takes place. With a some- 
what daring but illustrative comparison from modern medicine, one 
might say that, like a stomach the ulcers of which are deprived of 
their influence on the total organism by the severance of vagus and 
sympathicus and which is left to regulate itself, industry in post-capi- 

^* This — it has to be emphasized — holds for post-capitalist societies. It emphati- 
cally does not hold for industrializing countries, Communist or otherwise, where all 
the worst symptoms of European capitalism reappear. 

270 Classes in Post-Cafitalist Society (I) 

talist society has lost its all-embracing social significance and has to 
a certain extent been capsuled off from society. 

I shall not try here to trace in detail the historical genesis of the 
tendency described by these metaphors and comments. A number of 
trends could be indicated which have contributed to this development. 
T. H. Marshall sees one of the consequences of trade unionism in 
the creation of "a secondary system of industrial citizenship parallel 
with and supplementary to the system of political citizenship" (57, 
p. 44). If it is correct to speak of an "economic citizenship" side by 
side with the "political citizenship" (a phrase which trade unionists 
like to use), then this very distinction hints at the separation of the 
two spheres of industry and society. Industry appears as a society 
within society, a structural unit sui generis ^ which in a way is complete 
in itself without transcending its limits and without overlapping other 
structural units and associations. By way of illustration, we may also 
adduce here the much-discussed fact that the proportion of the pop- 
ulations of post-capitalist societies occupied in industrial production 
has not only failed to increase in the last decades but has, on the con- 
trary, decreased. The extension of "tertiary industries" plays a part- 
here. Even in the economy, industry is no longer that dynamic in- 
stitutional sphere which draws ever new groups into its whirlpool j 
here, too, a certain delimitation and stabilization can be demonstrated. 
The same tendency can, finally, be shown — if Sternberg's data and 
conclusions are correct — even for the volume of industrial produc- 
tion, that is, insofar as one can speak in Europe since the First World 
War and in the United States since the Great Crash, of a "halt in 
capitalist expansion" (see 186, pp. 177 ff.). 

The theory of the institutional isolation of industry in post-capi- 
talist society, if formulated on as general a level as we have so far 
done, gives rise to a number of obvious objections. We shall mention 
only a few. First, it would seem that the life of people in modern 
societies is increasingly rationalized, mechanized, in this sense "indus- 
trialized." The products of industry dominate consumption in all 
social strata and mold the standard and style of living of all. Second- 
ly, industrial and political problems are more closely connected than 
ever. If industrial production is stopped (e.g., by a strike) every 
member of society feels the consequences. In this sense, industry is 
by no means isolated. Thirdly, we can observe an increase rather than 
a decrease of government influence on industry. From this point of 
view, industry and society are also more closely connected than ever. 
Finally, there is even now a wealth of evidence of the existence of 

classes in Post-Capitalist Society (I) 271 

interrelations, often of a personal kind, between the ruling classes 
of industry and society. The composition of recent or current gov- 
ernments in the United States, Britain, France, Germany, and other 
countries would seem to sustain the thesis of C. W. Mills that the 
"power elite" of post-capitalist societies is relatively uniform and 
dominated by the carriers of industrial authority. It is not my in- 
tention to ridicule or underestimate objections of this kind. The least 
they prove is that the institutional isolation of industry is but a tend- 
ency and not a completed development. Yet, I believe that its ten- 
dential reality can be demonstrated despite all objections. In trying 
to present this demonstration, however, we shall do well to leave the 
investigation of the wider thesis of the isolation of industry to a 
separate study and confine ourselves to the more limited issue of the 
institutional isolation of industrial conflict. 

The thesis that, with industry, industrial conflict has been insti- 
tutionally isolated in post-capitalist society can be rendered more 
precise in a number of specific propositions, (i) We assert, in the 
first place, that position in the authority structure of industry and 
position in the authority structure of society are no longer necessarily 
identical, and that the industrial position of a man does not prejudice 
his political position. (2) It follows from this that neither latent 
nor manifest class interests of industrial conflict must be identical with 
the latent and manifest interests of the same people in political con- 
flict. Industrial interests have reference to the sphere of industry 
onlyj they aim, above all, at the maintenance or change of the in- 
dustrial status quo, not the inclusive social status quo. (3) It follows, 
further, that the dominant and subjected classes of industry need no 
longer be part of the corresponding political classes. The theory of 
classes permits the conclusion that there are as many discrete domi- 
nant and subjected classes in a society as there are associations. Here 
we assert that in post-capitalist society the rulers and the ruled of 
industry and society are tendentially discrete groups. (4) This means 
that membership in an industrial class leaves open to which political 
class an individual belongs, since independent determinants and mech- 
anisms of allocation are effective in the associations of industry and 
political society. 

Before we try to explore some of the empirical consequences of 
the theory described by these propositions, one more general remark 
may be in place. One of the astonishing and annoying characteristics 
of recent sociological theories of class conflict is that they frequently 
arrive at diametrically opposed conclusions. Whereas Burnham speaks 

272 classes in Post-Cafitalist Society (I) 

of an undiminished, even intensified, presence of the class conflict 
between employers (as managers) and workers, Schelsky believes 
that the question of the significance of the "old" classes for the present 
has to be answered "unequivocally in the negative." These incom- 
patible conclusions cannot be argued away by ridiculing one or the 
other student of the matter. Rather, the incompatibility of these 
theories challenges the sociologist to develop a new and better theory. 
This is precisely what the theory of the institutional isolation of in- 
dustrial conflict affords. It supports, so to speak, Burnham with respect 
to the sphere of industry where the old conflict persists (although not, 
as we have seen, in an intensified form), and it confirms, at the same 
time, Schelsky's notion by disputing the binding force of industrial 
conflict for the inclusive society. In this way, the claim for general 
validity implicit in both theories is rejected. The "old" conflict per- 
sists j but its effects are limited to the institutional sphere of industry. 
Outside industry, in the political society, the prolongation of "capi- 
tal" and "wage labor" into "bourgeoisie" and "proletariat" in the 
Marxian sense no longer describes the dominant patterns of social 


As formulated so far, the theory of institutional isolation of in- 
dustrial conflict is a general orientation, a point of view rather than 
a scientific theory. Above all, it is not, as such, susceptible of refuta- 
tion by empirical test. One may argue for or against it — but such 
arguments can finally neither confirm nor reject its validity. The 
theory proves testable only if we succeed in deriving certain testable 
propositions from it, and in confronting them with known or accessi- 
ble facts. Some of the assumptions which may thus be derived from 
the theory of institutional isolation of industrial conflict, and which 
promise to further the process of our analysis, will now be formulated 
and, insofar as possible, at least illustrated by empirical data. 

(i) If it is correct that with industry itself industrial conflict has 
been institutionally isolated in post-capitalist societies, it follows that 
his occupational role has lost its comprehensive molding force for 
the social personality of the industrial worker, and that it determines 
only a limited sector of his social behavior. Even today, occupation 
is generally assumed to be the decisive social role of at least the man 
in industrial society. "A man's job — occupying nearly one-third of 
his daily life — is more than just a means of livelihood or an outlet 

classes in Post-Capitalist Society (I) 273 

for his creative energy j it is a vital influence on his existence even 
beyond working hours. His social position, his economic welfare, 
and even his daily habits are all determined by the kind of job he 
holds" (120, p. 41 1). While there can be no doubt that income and 
prestige are in all industrial societies to a definable extent functions of 
occupation, and that many daily habits, such as the time of rising and 
going to bed, are determined by the requirements of the job, the 
thesis is advanced here that in the special case of industrial workers" 
that sector of social behavior which is not immediately determined 
by occupation is extending steadily. In post-capitalist society, the 
worker, when he passes the factory gate, increasingly leaves his occu- 
pational role behind him with the machines and his work clothes j 
outside, he plays new roles defined by factors other than his occupa- 
tion. The occupation and the expectations connected with it dominate 
less and less the life of the industrial worker, and other expectations 
mold his social personality. The job has become exactly what it is 
denied in the above quotation: a means for the end of life which as 
such assumes a fixed and limited place in the workers' behavior. 

Here, the "convergence of social science in highly industrialized 
countries" emphasized by Schelsky has its place j it aims at "putting 
the structures of consumption and leisure-time behavior into the cen- 
ter of the interpretation of contemporary society" (72, p. 65)- One 
may (and must) doubt whether consumption roles are becoming — 
as Schelsky thinks (72, pp. 6$ ff.) — "the central determinant of all 
patterns of behavior instead of class status," I should think that the 
undoubted trend toward increasing the importance of his consumer's 
position in the life of the worker in post-capitalist society develops 
at the expense of occupational position and testifies to the isolation 
of occupationally determined patterns of behavior, i.e., their restraint 
to a specific context confined in time and space. If it is possible to 
show, by empirical investigation, that consumption status does indeed 
assume a more prominent place in the consciousness and factual be- 
havior of the worker than his occupational status, the proposition 
formulated here may be regarded as confirmed. 

(2) A special case of the matter formulated in this proposition 
and a further consequence of the theory of the institutional isolation 
of industrial conflict can be found in the hypothesis that the partici- 
pants of industry, upon leaving the factory gate, leave behind them 

^'^ Probably this thesis is also valid for other occupational groups; however, jn the 
present context 1 want to restrict it to industrial workers. 

274 Classes in Post-CafitaUst Society (I) 

with their occupational role their industrial class interests also. The 
manifest contents of industrial class interests are no longer identical 
with those of political class interests. At a time when the one and 
only manifest interest of those subjected to industrial authority was 
an economic subsistence minimum, defined either physiologically or 
socially, this interest accompanied the workers to all spheres of their 
social behavior. But the significance for class conflict of the tendency 
toward equality, toward leveling the differences between strata, lies 
in the fact that the increasing equalization of living conditions and 
the institutionalization of the right to a minimum wage have provided 
a basis for separating the manifest interests of industrial and political 
conflict. Two aspects of this tendency are clearly visible today. First, 
the issues of industrial conflict in post-capitalist society are typically 
no longer of the kind that would divide the whole society into "two 
large hostile camps." They divide only the participants of industry 
into two hostile camps j they are confined to goals in which only these 
are "interested" by virtue of their occupational and, more specifically, 
authority roles. This is true not only of wage claims based on higher 
productivity, but also for demands to prolong paid vacations, shorten 
working hours, establish participation in management — to say nothing 
of the more technical issues of conflict (job evaluation, problems of 
time and motion study, of wage system, etc.) which, outside industry, 
are hardly understood as conflict issues. In terms of its issues, indus- 
trial conflict increasingly becomes industrial conflict without reference 
to general social and political problems. Secondly, this narrowing 
down of the issues of industrial conflict means that the individual 
worker is concerned with them only in his role as worker. In other 
roles he is moved by other things j as consumer or citizen, he is no 
longer worker. If, for the sake of clarity, an overstatement is per- 
mitted, one might say: as the bowling fan may get heated over cer- 
tain problems in his bowling club which leave his behavior outside 
the bowling club totally unaffected, so the issues of industrial conflict 
become increasingly irrelevant outside the sphere of industry.^® 

(3) One of the consequences of this fact and of the theory of 
institutional isolation is that industrial strikes no longer affect the 
so-called public of post-capitalist societies immediately; indeed, that 
strikes in one branch of industry leave workers in another branch rela- 

^^ Tendential assumptions always require qualification; this one is no exception. 
Obviously, matters of wage disputes affect the economic position of people; this in 
turn affects consumption chances and thereby the total social life. In historical per- 
spective, a separation of spheres seems apparent, however, even here. 

Classes in Post-Cafitalist Society (I) 275 

tively unaffected. If industry and society are two discrete social con- 
texts, conflicts in one do not lead to conflicts in the other, i.e., the public 
does not identify itself with the striking parties on the basis of the 
respective class situation, but for completely different motives (e.g., 
harm to the economy, to one^s own position, temporary discomforts, 
etc.). Geiger's "new line" of a conflict between those who participate 
immediately in production and the "mere consumers" plays a part 
here: events within the institutionally isolated sphere of industry 
may be regarded as threatening by the unconcerned public precisely 
because of this isolation; they are not identified with political con- 
flicts. It is a telling fact that in a German opinion survey in 1954 
almost one-half of those interviewed took a stand against strikes and 
expressed their belief that "the workers get, by strikes, advantages 
only for themselves and never even consider the interests of the 
general public" (see 229, p. 235).^^ 

(4) Within the sphere of political organization it follows from 
the theory of the institutional isolation of industrial conflict that trade 
unions and progressive (socialist, labor) parties are no longer iden- 
tical ; that, indeed, the notion of a workers' party has lost its political 
meaning. If it is true that industrial conflict is, in post-capitalist 
society, confined to its own specific sphere, if, therefore, the role 
distribution and the issues of industrial conflict have lost their em- 
bracing social significance, then the interest groups of industry and 
society are also discrete organizations. They may still be connected 
by the bond of tradition, but they are no longer united by a common 
cause and field of recruitment. The validity of this thesis is most 
evident in the United States and least evident in Britain. In all post- 
capitalist societies, however, there is the double tendency to establish 
the political independence of trade unions as interest groups in in- 
dustrial conflict and to extend socialist or progressive parties as interest 
groups in political conflict beyond the boundaries of an industrial class 
to "people's parties" or "mass parties." Many people are still sur- 
prised to find active unionists among the voters, or even functionaries, 
of conservative parties j but this is merely a consequence of the trend 
of development which we have tried to describe as one of institutional 

(5) If this theory proves a useful explanatory tool, it follows — 
to formulate a last derivation — that in post-capitalist society the ruling 

'■^ It has to be noted, however, that this latter attitude is probably more character- 
istic for Germany (where conflict is notoriously regarded as "bad" and undesirable) 
than for Britain and the United States. 

276 Classes in Post-Capkalisi Society (I) 

and the subjected classes of industry and of the political society are 
no longer identical j that there are, in other words, in principle two 
independent conflict fronts. Outside the enterprise, the manager may 
be a mere citizen, the worker a member of parliament; their industrial 
class position no longer determines their authority position in the 
political society. According to this assumption it is by no means true 
that "the relation to the instruments of production . . . decides the 
issue of class dominance, of power and privilege, in society" (Burn- 
ham, 140, p. 97) ; political authority is, rather, allocated independent 
of a man's industrial authority position. Moreover, this holds in- 
creasingly as within industry the separation of ownership and con- 
trol increases and as the more universal capitalists are replaced by 

It was the intention of this analysis to push the theory of institu- 
tional isolation of industrial conflict to the point at which testable 
hypotheses can be formulated. Since no quantity of empirical data, 
however large, is capable of verifying these hypotheses, it seems 
sensible to present them for criticism and, perhaps, refutation by 
empirical research. As long as they are not refuted, we may maintain 
that in post-capitalist society the associations of industry and society 
have to be understood as discrete universes of class conflict. "Wage 
labor" and "capital," the industrial classes of capitalist society, de- 
termine the social conflicts of industry even today as labor and man- 
agement, or trade unions and employers' associations, although the 
forms of these conflicts have been subject to many a change in the last 
hundred years. The fronts of social conflict, however, can no longer 
be extrapolated by merely extending the lines of industrial conflict 
beyond the boundaries of industry. Neither is "capital" extended 
into the "bourgeoisie," the ruling class, nor "wage labor" extended 
into the "proletariat," the subjected class, of post-capitalist society in 
its political aspect. Rather, this is a problem for separate investigation. 


The history of industrial conflict provides illustrations and exam- 
ples for most of the patterns and types suggested by the theory of 
group conflict in imperatively coordinated associations. These illus- 
trations and examples are all the more interesting, since the protago- 
nists of industrial conflict have in many ways remained the same 
throughout industrial development. We have seen that both capital 
and labor underwent a process of decomposition in the course of the 
last hundred years. The unified capitalist class of the time of Marx 

Classes in Post-Capitalist Society (I) 277 

dissolved into various elements, such as the managers, the mere 
owners, the finance capitalists j the homogeneous working class of a 
hundred years ago evolved within itself new lines of differentiation 
according to skill, income, prestige j both classes were extended and 
complicated by the emergence of a "new middle stratum" of bureau- 
crats and white-collar workers. However, all these developments 
have modified the authority structure of the industrial enterprise but 
slightly. There still are employers and employees, entrepreneurs 
and workers J and management and labor are even today the quasi- 
groups underlying the organizations that carry on the dynamics of 
the social development of industry. 

But the modes of their interrelations have changed in many ways 
since the time the Luddites in England fought their alienation by 
trying to break machines in isolated factories, or even since the time 
when a hired police force of management tried to disperse workers 
on strike by force. For one thing, the violence of industrial conflict 
has diminished considerably. This is due, in the first place, to the 
very fact of organization of the conflicting parties. Secondly, trade 
unions and employers' associations have established an often intricate 
system of routines of conflict regulation. Before they have recourse 
to violent manifestations of conflict, they meet to discuss their claims, 
they call in a mediator and, perhaps, an arbitrator} in short, they try 
to settle their disagreements by talking rather than fighting. This 
reduction of violence has been greatly helped by improvements in 
the standard of living of the workers. In most of the countries that 
might today be described as post-capitalist societies, absolute depri- 
vation on the part of industrial labor gave way, early in this century, 
to relative deprivation, so that the "cost" of victory or defeat in con- 
flict decreased. There is a peculiar dialectics in the fact that this re- 
duction of violence is both cause and effect of the institutionalization 
of industrial conflict: by forming organizations and defending their 
claims, management and labor have been able to introduce changes 
which in turn helped their chances of peacefully settling disputes. 

With the violence, although for different reasons, the intensity 
of industrial conflict has decreased also. Here, the dissociation of 
patterns that were superimposed in capitalist society is crucial. Above 
all, in contemporary societies industrial conflict and political conflict 
are no longer identical. The protagonists, issues, and patterns of in- 
dustrial conflict make for a discrete set of social relations. Industrial 
conflict has been severed from the antagonisms that divide political 
society} it is carried on in relative isolation. In many (though not all) 

278 Classes in Post-Cafitalist Society (I) 

countries, this dissociation of formerly superimposed patterns has also 
occurred with respect to the relations between class conflict and other 
types of conflict,^" and with respect to the relations between authority 
position and social status. Characteristically, the social status of a 
skilled man devoid of authority is higher than that of a low-level 
bureaucrat who participates, if in a distant way, in the exercise of 
managerial authority. We have not here analyzed the incidence of 
social mobility between the classes in industry. However, what little 
data we have^^ suggest strongly that there is a great deal of exchange 
between the classes in industry, between, as well as within, genera- 
tions. Industry is, of course, by no means "classless," but there is 
enough mobility to suggest that from this point of view, too, the 
intensity of conflict has diminished rather than increased. The con- 
clusion that the energies invested in industrial conflict by manage- 
ment and labor in post-capitalist society are rather smaller than a 
hundred years ago is confirmed, moreover, by numerous surveys 
showing a steadily decreasing interest and participation in union affairs 
on the part of labor. 

Diminishing intensity and violence of conflict have their effect on 
the modes of structure change in industry. They would suggest that 
sudden as well as radical changes are largely absent on the contempo- 
rary scene. Changes of conditions and structures occur gradually and 
remain piecemeal. In part, they take the form of penetration of the 
ruling groups by members of the subjected groups, but more often 
what happens in industry is that interests of labor are (if grudgingly) 
accepted by management and made part of the existing structure. 

In the foregoing analysis, I have confined myself to indicating 
such developments as point in the direction of decreasing intensity 
and violence. As always in social affairs, however, development is by 
no means unilinear. There are undoubtedly countertrends, also, and 
our analysis is not intended to suggest that all trouble in industry is 
past. For one thing, it is never possible simply to extrapolate social 
developments. The fact that industrial conflict has become less vio- 

^^ This is still not entirely true in Britain with respect to the position of the Irish 
in industry, and above all in the United States with respect to botli colored workers 
and immigrant worlcers. In these cases the management-labor conflict is even today 
often intensified by other lines of division. 

'^ Cf., above all, the data presented by G. Thomas for Britain (123, p. 30), ac- 
cording to which only 5 per cent of the "managerial" category in his representative na- 
tional sample have been managers all their lives, whereas no less than two-thirds of all 
managers have been in manual laboring occupations at some time. 

classes in Post-Capitalist Society (I) 279 

lent and intense in the last century does not justify the inference that 
it will continue to do so. On the contrary, experience shows that in 
the history of specific conflicts more and less violent, more and less 
intense periods follow each other in unpredictable rhythms. It is cer- 
tainly conceivable that the future has more intense and violent con- 
flicts in store. 

To some extent there are already indications of such a develop- 
ment. I have tried to show that not all changes introduced in order 
to regulate industrial conflict have been well-conceived. The German 
shop council and co-determination systems block rather than open 
channels of expression of conflicting interests. It would not be sur- 
prising if these ineffective modes of regulation should lead to new 
and at first uncontrolled outbrealcs of violence} indeed, it would, from 
the point of view of conflict theory, be surprising if this did not hap- 
pen. Moreover, the institutionalization of trade unions brings with 
it that phenomenon which Michels (see 200) so aptly described as 
the "iron law of oligarchy," i.e., the stabilization of an unequal au- 
thority distribution within the unions. This, again, promotes a new 
type of conflict, intra-union conflict, and one for which wildcat strikes 
already provide some evidence. It is hard to see how trade unions 
propose to check this development. Finally, there is always, of course, 
the possibility that totalitarian forms of government will interfere 
with the industrial order and change its patterns of conflict. It will 
be evident to the reader that most of the analysis of the present chap- 
ter presupposes a democratic political system. Once this system is 
abolished, a new situation arises inside as well as outside industry — a 
system (as I shall try to show in the following chapter) that promotes 
extremely intense and violent conflicts. Whatever trends toward re- 
duction of violence and intensity of conflict there may be in the in- 
dustry of post-capitalist society, there are countertrends also, and it 
is hard, if not impossible, to derive predictions in that sphere from 
the analysis of this chapter. 


Classes in Post- Capitalist Society 
II: Political Conflict 


Few people, sociologists or otherwise, deny that there still are 
conflicts in industry. Wage claims, demands for a share in manage- 
ment, strikes, and lockouts are too clearly in evidence to be argued 
away. But there are many people who claim that so far as political 
life in post-capitalist society is concerned, issues and lines of division 
other than class have become far more important for an understand- 
ing of society. Where this is argued, half-truths are usually inter- 
mingled with untruths. It is said that the problems of industry no 
longer concern every citizen; this is true, but it does not mean that 
there are no longer issues that divide the political community. It is 
said that political parties have become more and more similar in their 
programs J this is true, but it does not mean that they have become 
interchangeable either from the point of view of the voter or from 
that of policy. It is said that people no longer respond to ideologies j 
this is true, but it does not mean that they have no divergent convic- 
tions. At the bottom of such arguments there is often a vague asser- 
tion to the effect that by comparison with their attitude during the 
golden or gruesome age of capitalism, the people have changed their 
attitudes, views, and outlook on society. It is said that people no 
longer look at society in terms of divisions and antagonisms but now 
judge everybody on his merits from the point of view of a happy 
cooperative whole; and that sociologists who insist on cleavages and 
conflict merely again stir up troubles that have just been overcome. 
This assertion, to be sure, is not a half-truth but an untruth. Since it 
refers to people's feelings, however, rather than to sociological in- 
terpretations of reality, it is both necessary and appropriate that we 
begin our analysis of political conflict in contemporary societies with 
a look at some studies that are explicitly concerned with ascertaining 
how people see society. Discussion of these studies will link the so- 
ciological aspects of our problem with its psychological dimension and 

classes in Post-Ca-pitalist Society (II) 281 

at the same time add some color and verisimilitude to statements 
which as assertions, assumptions, and hypotheses remain by necessity 
vague, abstract, and, perhaps, unconvincing. 

I have indicated earlier at which points the sociological analysis of 
conflict may be fruitfully and legitimately supplemented by (social) 
psychological data. I have then emphasized what I again affirm here, 
that psychological evidence does not and cannot provide a conclusive 
test for our theory. Primarily, our theory refers not to what people 
think but to what they doj and while it may often be difficult to sepa- 
rate the two, the validity of our theory is in no way dependent on 
whether a representative sample thinks it is valid. The data to be 
discussed in this section illustrate some points of our theory, they are 
suggestive of further problems, they help to round off the picture 
of conflict in post-capitalist society presented here, but they must not 
be understood as either confirming or refuting the theory of conflict 
or any of its derivations. Between the scientific explanation and the 
popular view of a phenomenon there remains a gap which is hard 
to bridge. 

With respect to the question of how people see society, we are 
today in the fortunate position of having at our disposal a number of 
competent studies which provide suggestive and comparable data for 
various countries. Of these, I have singled out four that seem to me 
of particular import: those of Centers in the United States (38), 
Popitz and associates in Germany (69), Willener in French Switzer- 
land (76), andHoggart in Britain (52). Although these studies have 
been conducted almost entirely independently of each other,^ their 
authors formulate the object of their research in surprisingly similar 
terms. Above all, there are two notions that come up in most of them, 
namely, those of "class" and of "images of society" {images de la 
societSy Gesellschaftsbilder) . These studies are, in other words, di- 
rectly to the point of our own investigation. Although similar in sub- 
ject, these four studies differ in their techniques of investigation and 
methods of interpretation. Centers presents and briefly interprets" 
the quantified findings of a (multiple-choice) questionnaire survey. 

^ Willener explicitly refers to Centers but not to Popitz and Hnggart. All others 
are apparently (and, considering the dates of publication, understandably) ignorant 
of one another. The studies appeared in the following sequence: Centers (1949), 
Hoggart, Willener, Popitz (all 1957). The co-authors of Popitz's study were H. P. 
Bahrdt, E. A. Jüres, and H. Kesting; but since the parts in question here have been 
written by Popitz alone, it seems justified to refer to the study by his name. 

^ I am neglecting, here, the sketchy and rather out-of-place theoretical introduc- 
tion and conclusion of Centers' work. Contrary to the author's belief, these have, in 
my opinion, virtually no relation to his data. 

282 Classes in Post-Cafitalist Society (II) 

Willener used a questionnaire also, but an open-ended onej his more 
varied findings are quantified and interpreted. The study of Popitz 
and his team is based on depth interviews following a general schema j 
in his interpretation, Popitz largely avoids quantitative conclusions 
and instead presents a brilliant discussion. Hoggart's study, finally, 
is of a different kind altogether. It is an outstanding "impressionistic" 
account of what the author himself calls "aspects of working-class 
life," based on intimate personal knowledge but evidently entirely 
qualitative. In all cases, we shall confine ourselves in the following 
discussion to presenting in a summary fashion a small selection of the 
findings and conclusions — those that are relevant for the progress 
of our own investigation. 

One of the first conclusions reached by these studies is that, by 
and large, people do have an image of society. In a sense, this is per- 
haps not very surprising. The need to assess one's place in the world 
is presumably an existential needj it includes the necessity to place 
oneself in the social universe of reference. At the same time, though 
in general there exists an image of society, this does not come about 
merely as a matter of course, nor is it easily or precisely defined for 
the individual. However vague or stereotyped this image may be, it 
requires some considerable effort of reflection and of dissociation from 
one's most personal sphere. Let us look at how a man who lacks such 
reflection views the world. Popitz reports the case of a roll-caster 
in a steelworks who, in his opinion, has no image of society: "(What 
do you think of technical progress? ) I do not think that far. I always 
say, after us the deluge. (How does technical progress come about? ) 
The foremen do that, and the engineers and employers. . . . (What 
does co-determination mean to you?) Co-determination? Has that 
been carried out? I have very little interest in that . . ." (69, p. 
227) . This is certainly a pathetic statement; but should we not expect 
to find it often? It would seem that in fact it is anything but com- 
monplace to find that people have by and large a fairly articulate 
view of the society in which they live. 

A second uniform finding of the four studies in question is that 
people's images of society differ, and that such differences as there 
are are not random. Popitz and Willener each distinguish six types 
of approach. Those of Willener are directly related to problems of 
social stratification and class structure. According to him, people see 
society in terms of: {a) socioeconomic categories, {b) socio-occupa- 
tional categories, (c) a dichotomy of dependence, {d) class struggle, 

Classes in Post-Capitalist Society (II) 283 

{e) social prestige, or (/) political categories (see 76, p. 153). Popitz's 
classification of types is more general. According to the dominant idea 
in different images of society he distinguishes: {a) static order, (^) 
progressive order, {c) dichotomy as collective fate, {d) dichotomy 
as individual conflict, ((?) reform of the social order, (/) class struggle 
(see 69, p. 233). In both studies, the (Marxist) notion of society 
torn by class strife was found to be of but minor importance: 10 per 
cent of those interviewed by Willener and i per cent (!) of those 
interviewed by Popitz still adhered to this idea.^ The data presented 
by Centers and Hoggart do not allow a comparison here. 

Willener and Popitz agree, however, that their six types may be 
further reduced to two basic images of society which underlie the 
empirical variety found. Here, Centers' work may be adduced also. 
As we have done in this study. Centers distinguishes strictly between 
social strata and social classes, between a "static" and a "dynamic" 
view of society (38, pp. 26 ff.). For Centers, as for us, this is a theo- 
retical decision. However, the studies of Popitz and Willener show 
that this decision is reproduced in the views different people have of 
society. "Strata (or levels)," states Willener, "imply the image of 
a continuity . . . whereas classes indicate antagonistic groups" (76, 
p. 206), and he adds that, on the basis of his findings relating to this 
distinction, "there can be no doubt that there are fundamental dif- 
ferences between the image of individuals situated at the top of the 
social scale and those situated at the bottom" (p. 208). More pre- 
cisely: "The 'inferior* categories of respondents predominantly re- 
spond in terms of social classes rather than strata j conversely, the 
respondents of 'superior' categories have a tendency of referring more 
frequently to strata than to classes" (p. 206). This conclusion bears 
a striking similarity to that of Popitz when he expresses the notions 
of "class" and "stratum" by the terms "dichotomy" and "hierarchy": 
"All workers with whom we have spoken and who develop an image 
of society in the sense of our definition at all see society as a dichotomy^ 
incontrovertible or subject to change, unbridgeable or susceptible 
of mediation by 'partnership.' ... By contrast, the white-collar man 
knows a 'top' that is above him, and a 'bottom' that is below him. He 
places himself in the middle, and develops a remarkably acute sense 

^ The reason for this otherwise surprising difference between Switzerland and 
Germany is that whereas Popitz characterizes by the "class struggle" image only those 
who profess strictly Marxist views, Willener also includes those who vaguely refer to 
"capital," "exploitation," etc. 

284 classes in Post-Ca'pitalist Society (II) 

of distinction and of social gradations. One may assume, therefore, 
that he sees society not as a dichotomy like the industrial worker, but 
as a hierarchy" (69, pp. 237, 242). 

In these views, our two models of society reappear. The conti- 
nuity of a hierarchical system of stratification represents order and 
integration. There may be problems and strains, but there are no 
deep cleavages in society thus conceived. By contrast, the antagonisms 
of a dichotomous structure of class evoke the ideas of conflict, dis- 
sensus, and coercion. But, in the empirical findings of Popitz and 
Willener, these views are not complementary approaches to the same 
object. To be sure, even here they are not contradictory. But they 
are held by different people. Those "above" visualize society as a 
comparatively ordered continuous hierarchy of positions; those "be- 
low" are, above all, struck by the gap between them and "the others." 
Several interpretations of this strange and important fact are possible. 
In terms of our theory of conflict, however, it would seem that the 
dominant groups of society express their comparative gratification 
with existing conditions inter alia by visualizing and describing these 
conditions as ordered and reasonable j subjected groups, on the other 
hand, tend to emphasize the cleavages that in their opinion account 
for the deprivations they feel. At least potentially, there is an ideo- 
logical element in the models of society distinguished by us and, ap- 
parently, by people generally. The integration model, the hierarchi- 
cal image, lends itself as an ideology of satisfaction and conservation j 
the coercion model, the dichotomous image, provides an expression 
for dissatisfaction and the wish to change the status quo. Even at a 
time at which revolutionary ideologies of the Marxist type have lost 
their grip on workers everywhere, there remains an image of society 
which, in its political consequences, is incompatible with the more har- 
monious image of those "above," whether they be called "capitalists," 
"ruling class," or even "middle class." 

Little need be added about the hierarchical image of society as 
expressed predominantly by middle-class people. It is, clearly, de- 
rived from the notion of a bureaucratic hierarchy in which everybody 
has his defined place both above and below others. The whole is an 
ordered, well-organized system in which one can rise but not fall, 
and which has an accepted, institutionalized scale of symbols, titles, 
statuses. Such conflicts as are recognized are individual, highly per- 
sonal conflicts j and all other cleavages are banned from consciousness 
as unpleasant also-realities. 

As against this, even the more stereotyped versions of the dichot- 

classes in Post-Cafitalist Society (II) 285 

omous image of society are rather more colorful. From the studies 
under discussion it would seem that there are, in many languages, 
simple but descriptive terms by which to characterize the two sides 
of the dichotomy: them and us in Britain, ceux qui sont en haut and 
en has in Switzerland (and, probably, in France), die da oben and 
wir hier unten in Germany — these are expressions that belong to the 
stock-in-trade of working-class language. This is, for example, what 
some of the workers interviewed by Popitz had to say about co-de- 
termination: " 'One can always corrupt the workers' representatives. 
Where there is money, there is power. And if the workers' repre- 
sentatives have really pushed through a decision, one can twist its 
execution in such a way that nothing comes of it. . . . They don't 
want others to look at their cards. ... If they ever go in for co- 
determination, then to their own advantage: for the worker as much 
co-responsibility and as little co-determination as possible.' . . . 'We 
have nothing to co-determine. That is determined by management. 
What they say, will be done — and that is that.' . . . 'All that is just 
talk. We have nothing to co-determine. It's been managed by the 
unions, and if anybody co-determines, it is at best the union secre- 
taries and bosses. . . . They up there don't care anyway' " (69, pp. 
202 f.). Hoggart summarized the attitude apparent from these state- 
ments in a splendid selection of stereotypes: " 'They' are 'the people 
at the top,' 'the higher-ups,' the people who give you your dole, call 
you up, tell you to go to war, fine you, made you split the family in 
the 'thirties to avoid a reduction in the Means Test allowance, 'get yer 
in the end,' 'aren't really to be trusted,' 'talk posh,' 'are all twisters 
really,' 'never tell yer owt' (e.g., about a relative in hospital), 'clap 
yer in clink,' 'will do y' down if they can,' 'summons yer,' 'are all in 
a click [clique] together,' 'treat y' like muck' " (52, p. 62). 

The criteria of distinction between the two groups that make up 
the dichotomy of society are varied. Hoggart's list of idioms carries 
a strong connotation of resentment of authority, so that differences 
of power might indeed appear to mark the dividing line between 
"them" and "us": " 'Them' is the world of the bosses, whether those 
bosses are private individuals or, as is increasingly the case today, 
public officials" (52, p. 62). Willener emphasizes the criterion of 
power also, but he adds others: "Certain interviewees conceive essen- 
tially two classes: the salaried and the nonsalaried, in other words 
those who are dependent and those who are independent. To this may 
be added the formula 'those who work and those who do not work' 
which is given more rarely with the same meaning" (76, p. 155). 

286 Classes in Post-Ca-pitalist Society (II) 

Popitz again stresses the importance of the "stereotype alternative 
of power and impotency" (69, p. 244), but regards another criterion 
as equally significant: the dichotomy between manual and nonmanual 
labor. This dichotomy is, for Popitz, closely related to the "formula" 
Willener encountered among his respondents : "Even very intelligent 
workers who make a point of judging white-collar people justly and 
who concede that there have to be such people, too, remain mistrustful 
in one respect: it seems extremely questionable to them whether white- 
collar people really work" (69, p. 2385 cf. 68). White-collar work 
lacks "publicity," it is not susceptible of the same kind of control as 
manual work,- it is not as visibly work; and a line is drawn between 
those who do "visible" and those who do "invisible" work. In this 
question of criteria for social dichotomies the study of Centers is also 
relevant, for, although he did not intend to do so. Centers also found, 
in effect, that for most people (American) society consists of but two 
classes: the middle class and the working class. Ninety-four per cent 
of his original interview population assigned themselves to these two 
classes.^ Among the criteria of distinction between these classes, 
Centers found "beliefs and attitudes," "family," and "money" most 
prominent. But, like Willener, Centers emphasizes: "To the mem- 
bers of the working class the most important criterion of middle class 
membership after money or income is the ownership of a small busi- 
ness, profession or trade 5 in sum, being an independent operator or 
proprietor of some kind" (38, p. 99). By contrast, "it cannot help 
but strike one as highly significant that the most distinctive criterion 
given for membership in [the working] class is 'working for a living' " 
(38, p. 100).^ Not surprisingly. Centers, much like Popitz and Wil- 
lener, concludes that "the effect is to make the white-collar work vs. 
manual work and salaried work vs. wage work dichotomies both im- 
portant bases for class distinction in virtue of their importance as 
criteria for working class affiliation, and the psychological effect is to 
push white-collar workers toward identification with the middle class" 
(38, p. 102). 

* In a later study Centers reports only 88 per cent did so, while there was a 
slightly larger proportion of interviewees that "didn't know" or ranked themselves as 
"upper" or "lower class," the four alternatives presented to them being "upper," 
"middle," "working," and "lower class" (see 38, p. 77). 

^ Even more striking, perhaps, is the comparative evidence which seems to agree 
even in the details of everyday language : "working for a living and not working for a liv- 
ing" (Centers, U.S.A.), "ceux qui travaillent et ceux qui ne travaillent pas" (Willener, 
Switzerland), "wirklich arbeiten und nicht wirklich arbeiten" (Popitz, Germany). 

Classes in Post-Cafitalist Society (II) 287 

However, all four authors stress the fact that these criteria of 
distinction do not permit of an entirely unequivocal identification of 
all persons and occupations as "middle class" or "working class," 
"above" or "below." The studies seem to agree that the dividing line 
between "them" and "us" runs somewhere through the group of 
salaried employees. "The class position of office workers seems also* 
to be an equivocal one, for they are claimed almost as often by work- 
ing class people as they are by the middle class. The difficulty of 
assigning them to class membership appears thus no less a one with 
the members of actual classes than it has been to social scientists" (38, 
pp. 81 f.). Willener thinks that by the "below" most people mean 
wage-earning as well as salaried employees (see 76, p. 163), but I 
think more probable Popitz's and Hoggart's suggestion that in peo- 
ple's minds the phenomenon of social distance is crucial for distin- 
guishing between "them" and "us." According to Popitz, the fore- 
men and immediate supervisors of industry, whose work is visible 
to most workers, are usually counted among the working class, and 
"above" begins with the shop supervisors and the shop councilors 
and union secretaries, for these no longer "belong" (see 69, pp. 
243 f.). According to Hoggart, for the worker the world of "them" 
begins even earlier: "So, when working-class people are asked to 
become foremen or N.C.O.'s they often hesitate. Whatever their 
motives, they will be regarded now as on the side of *them' " (52, 
p. 64). It will be well to remember, in the following analysis, that 
for those who visualize society as dichotomous, the upper part of the 
dichotomy begins not far from the bottom layer of social stratifica- 
tion and includes all those who have even a minimal share in the 
exercise of authority. 

Whatever changes may have occurred in the last hundred years, 
the idea that there is a fundamental division of society into "haves" 
and "have-nots," "above" and "below," "them" and "us" is still a 
force in the minds of many people. One might be tempted to think 
that the dichotomous image of society is a relic of Marxism or, more 
generally, of the conditions of early capitalism and their interpre- 
tation. However, Ossowski has shown that "the dichotomous view 
of social stratification" — as an essay of his is entitled (96) — is both 
older and more general than capitalism. "The spatial metaphor which 
represents society as an aggregate of men of which some are above 
and others below belongs to tITose images which do not lose their 

•Centers had just discussed the position of farmers. 

288 Classes in Post-Cafitalist Society (II) 

pertinence in the course of the centuries and which, as the history of 
cultures seems to prove, impose themselves to the imagination" (96, 
p. 16). Ossowski follows this image through the myths and religions 
of mankind, through literature and philosophy. Three aspects of the 
social dichotomy run like a thread through its many manifestations: 
the divisions into rulers and ruled; rich and poor; those for whom 
one works, and those who work (96, p. 19). We might feel inclined 
to replace, in the medieval English quatrain, the word "God" by 
"society" and thereby take the worst sting out of the dichotomy, but 
its substance remains essentially true even today: 

The rich man in his castle, 

The poor man at his gate, 
God made them high or lowly. 

And ordered their estate. 

The assertion that a dichotomous image of society is an archetype 
of human understanding cannot, of course, be supported by the evi- 
dence presented in this section. Rather, there would seem cause for 
qualifying the conclusions already drawn. We have seen earlier that 
not everybody has an image of society at all; there are people who get 
along with a minimum of reflection on matters beyond the immediate 
horizon.^ Of those who have an image of society, only a few visualize 
society as a dichotomous entity. If one breaks down people's views 
by their own occupational or class position, there are indications of 
"deviance" from the point of view of class theory: workers rank them- 
selves as "middle class," white-collar people profess a dichotomous 
image of society. There are problematic groups, groups that are not 
easily placed in terms of the dichotomy, such as salaried employees 
and farmers. Finally, the consequences people draw from their 
images of society differ greatly. Centers believes that "the top occu- 
pational strata are marked by their adherence to the status quo in the 
order of politico-economic relations. In contrast, the lowest occupa- 
tional groups are distinguished by their lack of support of the status 
quo and by their endorsement of views clearly radical in character" 
(38, p. 208). Plausible as this sounds, it is as much a direct derivation 
from Centers' theory as it is a summary of evidence. Very likely, 

^ Popitz found that about 20 per cent of those interviewed by him have no real 
image of society (6g, p. 233) ; 24 per cent of those interviewed by Willener responded 
negatively, without clear idea, or not at all (76, p. 161); by comparison, the 2 per 
cent who "didn't know" or "didn't believe in classes" in Centers' study are surpris- 
ingly few (38, p. 77). 

Classes in Post-Cafitalist Society (II) 289 

Popitz is closer to the truth in concluding that the "consciousness" 
of workers "may still admit to itself and to others that industrial 
workers have their own interests which are in conflict with those of 
others" (69, pp. 247 f.), but that there are also many phenomena, 
both social and personal, that make workers hesitant to draw pro- 
nouncedly radical political conclusions from their dichotomous image 
of society. This is confirmed by Glantz, who found in an independent 
empirical study in the United States that "a latent tendency towards 
radicalism undoubtedly exists among some workers, but there is little 
or no historical evidence to indicate that it has recently been develop- 
ing into a conscious ideology" (195, p. 378).® Ossowski has shown 
that the dichotomous image of society may be activated into an ideol- 
ogy of political conflict, but it is as such no more than an interpreta- 
tion of the social world. 

As a way of seeing society, however, the dichotomous view is a 
solid and, probably, powerful social fact. It may lend itself toward 
giving additional force to the considerations offered in this study. 
In a way, it provides a second foothold for our analysis of post-capi- 
talist society: on the one hand, there are the suggestions and assump- 
tions derived from the theory of conflict j on the other hand, there are 
indications and problems furnished by the findings of systematic em- 
pirical observation. In any case, these findings dispose of the half- 
truths and untruths mentioned at the outset of this section. That from 
one point of view society presents a dichotomous image, an image of 
conflict and dissensus, is by no means an invention of sociologists 
imposed on a basically harmonious and cooperative social reality. The 
dichotomies of post-capitalist society may have little to do with those 
asserted by Marx: there is no disagreement here between sociological 
and public opinion. But there still are dichotomies, and they are very 
real to those who experience society in terms of them. Here, as else- 
where, sociological analysis is more than a disengaged and noncom- 
mittal exercise of the mind: it is an attempt to explain rationally and 
systematically those facts which to men in society are real stumbling 
blocks on their path. 


"An imperatively coordinated association will be called political 
association if and so far as the enforcement of its order is carried out 

® Glantz is explicitly — and rightly — critical of Centers' conservatism-radicalism 
scale and characterizes most of the conclusions based on this scale as "methodologically 
derived fiction" (195, p. 378 n.). 

290 Classes in Post-Cafitalist Society (II) 

continually within a given territorial area by the application and threat 
of physical force on the part of the administrative staiff'^ (Weber, 
33(3, p. 1 54) . More precisely, Weber defines the state by its monopoly 
of power within a given territory. Both these definitions are not 
altogether without their difficulties. At first sight it would seem that 
the state, conceived as an imperatively coordinated association, is 
inclusive of all other associations of society, since every citizen has 
a place within it. In other words, whereas industrial enterprises, 
churches, and armies are clearly delimited from each other and from 
other associations by a definite personnel, the state is an aspect of the 
total society which, it might appear, is impossible to isolate and to 
contrast with associations "within it." However, this appearance is 
demonstrably deceptive. It is inaccurate to say that all men in a given 
territorial area belong to the state : they do so only as incumbents of 
one position and role, namely, that of citizen. Strictly speaking, the 
association of the political state is an organization of roles. The em- 
pirical priority of its structure and dynamics cannot be derived from 
an assumption that it is, by definition or theory, more inclusive than 
other such organizations. What some social scientists refer to as the 
"polity" is a unit of social analysis strictly equivalent to an industrial 
enterprise, a church, or an army. For each of these associations we 
may ask how its structure generates conflicts, how these conflicts are 
patterned, and what changes result from them. 

Thus the state, like the enterprise, has a structure of authority 
relations. There are in it positions and persons endowed with the 
right to issue authoritative commands (laws, decrees, etc.), and there 
are other positions and persons subjected to them. The state "is an 
association like others: churches, trade unions, and the rest. It differs 
from them in that membership is compulsory upon all who live within 
its territorial ambit, and that it can, in the last resort, enforce its obli- 
gations upon its subjects. But its moral character is no different from 
that of any other association. It exacts loyalty upon the same grim 
condition that a man exacts loyalty from his friends" (197, p. 37). 
The important point to emphasize here is that in the state, as in in- 
dustry, authority is exercised by certain persons by virtue of their 
positions. There has been a considerable amount of opinion in recent 
years to the effect that in post-capitalist society political authority, as 
it existed in most or all earlier forms of social organization, has been 
"depersonalized" and in that sense abolished. In interpreting the 
findings of Popitz reported above, Schelsky states: "But the substance 
of this dualistic understanding of society begins to change markedly: 

classes in Post-Capitalist Society (II) 291 

whereas *we^ means more and more clearly 'those who do the same 
work with me here,' the opposite pole is increasingly identified not 
with capitalists or the bourgeois, but with the anonymous forces of 
all kinds of large bureaucratic organization and its functionaries" (72, 
pp. 67 f.). The fallacy implicit in this statement reminds one both 
of Renner and of Riesman, who, each in his own way, also assert the 
disappearance of traditional authority. As we have seen earlier, Ren- 
ner believes that in post-capitalist society the rule of the few has been 
replaced by the rule of "the law" or the "general will": "This gen- 
eral will defines the goal of society and of the economy, and all func- 
tionaries pass from the service of masters to the service of the whole. 
This general will is the law." Of course, laws have to be executed, 
but this is a "matter of institutions which are organized exclusively 
according to economic-technical standards" (71, p. 227). In other 
words, "objective" forces have taken the place of ruling groups. 
Riesman's conclusions are a little more cautious. But when he speaks 
of the "thinning extremes of those who were once leaders and led" 
(230, p. 248), when he refers to the "amorphous distribution of 
power" in contemporary (American) society, when he asserts that 
"while it may take leadership to start things running, or to stop them, 
very little leadership is needed once things are under way" (p. 252), 
he is not far from the theory that the personalized authority of a 
ruling class is a thing of the past. "Power in America seems to me 
situational and mercurial; it resists attempts to locate it the way a 
molecule, under the Heisenberg principle, resists attempts simulta- 
neously to locate it and time its velocity" (p. 252). There is a trace 
of Hegelianism in the conceptions of Schelsky, Renner, and Riesman, 
and we shall have to reject them. 

In the first place, it is not very prudent to render the analysis of 
problems unnecessarily difBcult by destroying the tools which are 
necessary for its successful accomplishment. If we want to investi- 
gate changes in the distribution and character of political authority, 
we do well to hold on to this concept. "The modern state, for prac- 
tical purposes, consists of a relatively small number of persons who 
issue and execute orders which affect a large number in whom they 
are themselves included" (197, p. 295). The important clause in 
this definition of Laski's is that it is useful "for practical purposes." 
Laski was well aware that "a view such as this has at least the supreme 
merit of realism. It admits that acts emanate from persons, and it 
insists that those persons are subject to the scrutiny of their fellow- 
citizens" (p. 36). Even in order merely to describe problems con- 

292 Classes in Post-Cafitalist Society (II) 

nected with the authority structure of the state, we have to be able 
to identify this structure as one of super- and subordination of per- 
sons in their capacity as incumbents of social positions. 

This approach seems realistic in a second sense, too. The state- 
ment that the force of the law has replaced the authority of persons 
provokes the question: Why was there this difference in other socie- 
ties? The notions of "anonymous" ruling forces or of a society that 
runs itself seem to document no more than that it is difficult, or that 
Schelsky and Riesman find it difficult, to identify the seat of political 
authority in post-capitalist societies. But does this mean that there 
is no authority? Does it mean that there are no persons who, by virtue 
of their positions, are entitled and expected to make authoritative 
decisions? It seems to me that a Hegelian reification of the state as 
liberty incarnate is required in order to answer these questions in the 
affirmative. Can there be any doubt that there still are governments, 
parliaments, and courts of law in the contemporary world? And — 
since this is obviously an entirely rhetorical question — can there be 
any doubt that cabinet ministers, members of parliament, and judges 
lend their authority to decisions that affect the lives of those subjected 
to them? If there is any truth in attempts to dispute the continued 
presence of political authority in the contemporary world, it is in the 
empirical generalization that many people find it hard to name and, 
perhaps, identify those "above." But surely it is not the task of the 
sociologist to hypostasize the confusions of public opinion into ambi- 
guities of social structure itself. 

However we may eventually describe and delimit the ruling class 
of the polity of post-capitalist society, we maintain that the presence 
of an unequal distribution of political authority over persons as in- 
cumbents of positions is both a useful assumption and a descriptive 
fact. Furthermore, most of the elements of the authority structure 
of the polity are easily identified and clearly visible to anybody. There 
is, in the first place, the large quasi-group of those who have no share 
in the exercise of political authority. They may be described as the 
"mere" citizens, i.e., those who occupy no political position other than 
that common to all members of the polity." It is a characteristic fea- 

® Strictly speaking, citizenship is of course dependent on more conditions than 
mere residence in a given territory. Children, lunatics, criminals, in some countries 
women, new immigrants, and certain other groups do not enjoy citizenship rights. 
In modern democratic societies we can neglect these groups in an analysis of political 
conflict; in some earlier societies, however, such as classical Athens, the very possession 
of citizenship rights involved political authority, and the dividing line between domi- 
nance and subjection could be defined by possession or nonpossession of citizenship 

Classes in Post-Capitalist Society (II) 293 

ture of political conflict in post-capitalist society that even the "bot- 
tom" of its authority structure is not entirely deprived of rights. A 
citizen is not only eligible for political office, he has, above all, the 
right to votcj and the act of voting may be described as an exercise 
of control over others. Legitimacy was, in all societies, a necessary 
requisite of the exercise of authority. In modern democracies the pre- 
sumption of legitimacy has been converted into a continuous process 
of legitimation through regular elections and, in some cases, plebi- 
scites. It may therefore be argued that no citizen of a democratic state 
is entirely powerless with respect to its political affairs. However, 
despite this basic power common to all, a clear line can be drawn 
between those who enjoy nothing but this minimum and those who 
are in the position to exercise regularly control over the life chances 
of others by issuing authoritative decisions. The citizens of a demo- 
cratic state are not a suppressed class, but they are a subjected class, 
or quasi-group, and as such they constitute the dynamic element in 
political conflict." 

By contrast to "mere" citizens, members of the three classical 
branches of government have authority and, therefore, constitute the 
quasi-group of those in domination." There is, first, the legislative 
branch. De jure — if not de facto — it is, in most present-day advanced 
societies, embodied in parliaments, chambers of deputies, houses of 
representatives, and the like. Deputies, representatives, members of 
parliament belong, by virtue of their position, to the ruling quasi- 
group of the polity. Strictly speaking, however, this does not hold 
for all members of parliament in a given situation: only those mem- 
bers represent a more or less permanent part of the authority struc- 
ture of the state who belong either to the majority party or to those 
parties which make up a governing coalition. Opposition members 
also have some degree of authority by virtue of their parliamentary 
position J they may sanction decisions of the government in power, 
assent to, or even inaugurate legislation and thereby exercise authori- 
tative control J but theirs is an authority that might legitimately be 

^° Laski, in his definition of the state quoted above, emphasizes that those who 
execute orders are themselves affected by them. This important clause indicates a 
peculiarity of the constitutional (as against the absolute) state: in it, even the incum- 
bents of positions of domination are at the same time citizens. They have authority 
only in one of their (political) roles; by this duplicity of roles they control themselves, 
so to speak. 

■^^ This statement will presendy be qualified in two respects: first, by the assess- 
ment of the class position of political bureaucracies, and, second, by the discussion of 
the phenomenon of representation, or the relation between power and interests in the 
modern state. 

294 Classes in Post-Capitalist Society (II) 

described as "situational." Members of the parliamentary opposition 
exercise authority only if and when they are in agreement with the 
majority or convince the majority to agree with them. Otherwise, 
they represent the interests of those excluded from domination. By 
contrast to members of the parliamentary majority, they do not, by 
virtue of their positions, permanently belong to the ruling quasi- 
group of the polity. Obviously, this subtle distinction is blurred in 
reality by relations of personal influence, complications of specific sit- 
uations,^^ and, above all, by the implications of a representative system 
in which group interests are regularly, and sometimes institutionally, 
represented by nonparliamentary organizations. We shall turn to 
these and other complications presently. 

As far as the judiciary branch of government is concerned, de- 
limitation of "mere" citizens and carriers of authority presents least 
difficulty, although here, too, a "situational" element of the exercise 
of authority is not absent. In the first place, the judiciary consists 
of all positions and persons who are permanently in the service of 
the state in order to enforce, and sometimes modify, the law: legal 
advisers, judges on all levels, public prosecutors. To these are added 
the two groups — solicitors and juries — whose authority is confined to 
specific situations and who cannot therefore as such be included in 
the ruling quasi-group of the polity. To refer to the judiciary as 
part of the ruling class is of course a somewhat doubtful way to express 
this idea. I do not mean to imply a Marxian view of the legal system 
as an incarnation of the ideology of the ruling class. Rather, the 
position of the judiciary would seem to be somewhat similar to that 
of the bureaucracy in that both are branches of government without 
being the seat of particular interests except for a vague general orien- 
tation to uphold the status quo. In other words, the judiciary need 
not be conceived of as a militant part of the dominant conflict group 
of the polity, but its members are in the position to exercise control 
over others. They are therefore placed on the plus-side of the zero- 
sum distribution of authority in the state. 

Finally, the executive branch represents an integral part of the 
dominant positions in the political state. Executive positions are, 
above all, government positions. Cabinet ministers and secretaries of 
state constitute the visible expression of the chief executive, and by 

^^ Such as, e.g., very small majorities dependent on the attendance of members, 
or the independence of legislative and executive in the American sense which makes 
a contradictory balance of power in the two branches possible. 

Classes in Posi-Capitalist Society (II) 295 

virtue of this fact they are rightly considered the real exponents of 
political authority. But the executive complex of the modern state 
extends beyond the members of a cabinet. It includes a large num- 
ber of positions which might be described as those of white-collar 
employees. And it is these positions that require our special atten- 
tion in an analysis of political conflict in post-capitalist society. 


Even in his definition of the political association, Max Weber 
referred to the "administrative staff" as executor and instrument of 
authority. An administrative staff can today be found in all three 
branches of government. There are bureaucrats associated with the 
legislative, with the judiciary, and with the executive, although their 
number is probably greatest in the various branches of the executive. 
Bureaucracies are moreover characteristic not only of the political 
system but also of all other institutional orders and associations of 
post-capitalist society. We have already seen that the rapid growth 
of an industrial bureaucracy introduced new elements into the pat- 
terns of conflict in industry. The analysis of the class position of 
bureaucracies presented in this section applies, therefore, to the ad- 
ministrative staff of all associations; its immediate reference, how- 
ever, is to the authority structure of the political state. 

It is perhaps appropriate to recall at this point that in our survey 
of historical developments in the last hundred years we have divided 
the large group usually referred to as "white collar" or "new middle 
class" into two distinct parts. One of these, that of white-collar 
workers, clearly belongs to the subjected quasi-groups of imperatively 
coordinated associations. There are of course white-collar workers in 
the service of the state also. In many countries, railway and post-office 
employees, workers in public enterprises and municipal utilities, etc., 
have salaried or even civil service status. The following discussion, 
however, is concerned not with these occupations and persons whose 
class position is unambiguous, but with bureaucrats of all levels, i.e., 
with those salaried white-collar employees whose position places them 
on a step of the ladder of administrative jobs. What is their position 
in political conflict? Are they, as administrators of the means of in- 
dustrial authority, a part or even the whole of the ruling class? Or 
are they, as dependent, often subordinate salaried employees, an ele- 
ment of the subjected class? There are few subjects which have been 
dealt with as thoroughly and extensively by sociologists as that of 

296 classes in Post-Cafitalist Society (II) 

bureaucracy. True, most investigations are not explicitly related to 
the problem of class, but they are no less useful in our context for this 
restriction. More than elsewhere in this study we can fall back, here, 
on studies by other authors and claim empirical verisimilitude for our 

Bureaucratic organizations differ from industrial organizations 
in one important point. Whereas the authority structure of industrial 
organizations ifso facto defines the borderline that divides the two 
aggregates of those in positions of dominance and those in positions 
of subjection, and whereas industrial organizations are in this sense 
dichotomous, bureaucratic organizations typically display continuous 
gradations of competence and authority and are hierarchical. Within 
dichotomous organizations class conflict is possible j within hierarchi- 
cal organizations it is not. This difference has an important conse- 
quence for the definition of bureaucratic roles. Insofar as bureau- 
cratic roles are defined in the context of a career hierarchy, they do 
not generate a (class) conflict of interests with other bureaucratic roles. 
Tensions within bureaucratic role structures arise characteristically 
from frustrations of expected opportunities, from "blocked or *once- 
for-alP mobility" (Tropp, 132, p. 322). If, for example, there are 
in a bureaucratic hierarchy several points of entry which divide the 
total career into subsections separated by insurmountable barriers, 
these tensions may of course lead to social conflicts also, i.e., to con- 
flicts between groups. However, one of the distinctive features of 
bureaucratic role structures is the principle of competition between 
individuals, and the psychological strains and tensions resulting from 
it (cf. Merton, 136). Structural conflicts based on latent antagonisms 
of interest in the sense of class theory are absent from bureaucratic 
organizations by virtue of their hierarchical character. This means 
that all incumbents of bureaucratic roles in the association of political 
society belong on the same side of the fence that divides the positions 
of dominance from those of subjection. Since a break such as gen- 
erates class conflict cannot be demonstrated within bureaucratic struc- 
tures, it follows that bureaucracies stand in their entirety either on 
the side of dominance or on that of subjection — no matter how dif- 
ferentiated bureaucratic roles may be according to qualification, status, 
and sphere of authority. 

^^ The following discussion has been inspired largely by the works of Weber (33), 
Bendix (126 and 138), Merton (136), Lockwood (135), Gouldner (131), and 
Croner (129), as well as by a number of further essays brought together in the reader 
edited by Merton and others (136). 

classes in Post-Ca-pitalist Society (II) 297 

It is hardly necessary to emphasize that it would be nonsensical 
to describe the roles of political bureaucracy in their entirety as roles 
of subjection. At least the highest civil servants are unequivocally 
bearers of political authority; and what is true for the highest civil 
servants is also true — in accordance with the premise just introduced 
— for all other employees of political administration. Bureaucratic 
roles are roles of political dominance. Their definition includes cer- 
tain latent interests which aim at the maintenance of existing insti- 
tutions and valid values. The incumbents of bureaucratic roles are 
members of one and the same ruling quasi-group. At this point the 
analytical value of a rigidly formulated "theory of delegation" be- 
comes apparent: the development which Weber described as one of 
" 'socialization* of authority relations" (33^', p. 669) has made for 
a "division of labor" of the total process of the exercise of authority. 
Like the division of labor in industrial production, this has led to the 
creation of numerous specialist positions, every one of which bears 
but slight traces of the process of which it is a part. Who produces 
the car in an automobile factory? The director? The fitter? The 
foreman? The typist? Every one of these questions has to be an- 
swered in the negative, and one might therefore be tempted to con- 
clude that nobody produces the car at all. Yet the car is being pro- 
duced, and we can certainly identify people who do not participate 
in its production. In relation to authority, both industrial and political, 
we encounter a strictly analogous situation. Nobody in particular 
seems to exercise "the authority," and yet authority is exercised, 
and we can identify people who do not participate in its exercise. 
Thus the superficial impression of subordination in many minor 
bureaucratic roles must not deceive us. All bureaucratic roles are 
defined with reference to the total process of the exercise of authority 
to which they contribute to whatever small extent. Differentiation 
of spheres of authority (i.e., of how many people are controlled to 
what extent) permits of manifold gradations j but the possession of 
authority itself does not. The strange structural state emerging for 
these reasons in post-capitalist society has been described by Weber 
as a "leveling of the ruled vis-a-vis the ruling, bureaucratically dif- 
ferentiated group" (33^, p. 667). While the "mere" citizens, those 
excluded from political authority, are in this respect a uniform 
(quasi-) group, the ruling class presents an image of far-reaching 
hierarchical differentiation. 

It is clear that this type of analysis of the class position of clerks 
and bureaucrats confirms the subjective experience of people described 

298 Classes in Post-Capitalist Socieiy (II) 

earlier in this chapter, where we have seen that for many "ordinary" 
citizens and workers the world of "above" or "them" begins with the 
white-collar employee. " 'Being dressed up to go to work,' 'sitting 
down to work,' 'having it cushy,' 'pushing pens,' 'being bosses' men,' 
'being one of them up there* from whom orders come and authority 
emanates, are terms which working men frequently use of office 
workers, irrespective of their actual positions in the office hier- 
archy" (Lockwood, 135, p. 131). The fact that white-collar people 
do not usually regard themselves as being "on top" or "upper 
class," far from contradicting the attitude of those "below," adds a 
further element which we have here discussed: the very nature of 
bureaucratic organization makes everybody feel that although some 
are "below" him, there are also others "above," so that he is "in the 
middle." Whatever the ruling class of post-capitalist society may be, 
it seems probable that it will tend to deny its own rule on account of 
its internal differentiation, which places every one of its members 
between two others. However, while this internal differentiation of 
the ruling class complicates the situation, sociological analysis and 
people's "class consciousness" converge in the conclusion that "the 
clerk is the man on the other side of the desk who is somehow asso- 
ciated with authority" (Lockwood, 135, p. 132). 

According to Weber's conception — which, more recently, has been 
adopted and further elaborated by Bendix — bureaucracy has not only 
a share but a monopoly of political authority. In support of this thesis, 
both sociologists refer, above all, to the monopoly of specialized 
expert knowledge on the part of bureaucracies which makes their re- 
placement (say, in a revolution) as impossible as their effective control 
by others who invariably find themselves, "in face of the schooled 
civil servant who stands in the process of administration, in the posi- 
tion of the 'dilettante' vis-a-vis the 'expert' " (Weber, 33^, p. 671). 
This monopoly position is further strengthened by the attributes 
characteristic of bureaucratic status, and by tenure^ pension rights, and 
similar privileges. It would be misleading to think of bureaucracies 
as a ruling caste because of their monopoly of authority j a class be- 
comes a caste only if entry to it remains, for many generations, an 
exclusive privilege of the children of its members. This is decidedly 
not the case in highly mobile advanced industrial societies. Never- 
theless, the monopoly of authority founded in expert knowledge pro- 
vides the incumbents of bureaucratic roles at any given time with a 
degree of exclusiveness which underlines their unity as members of 

classes in Post-Cafitalist Society (II) 299 

the same quasi-group and may be conceived of as a condition of con- 
scious solidarity. 

Does all this mean that the incumbents of bureaucratic roles in 
public administration are the ruling class of post-capitalist society? 
The attempt to answer this question reveals in the class position of 
bureaucracy a strange paradox which explains many otherwise in- 
comprehensible modes of behavior of the clerk, the administrator, 
and the civil servant. "The fact that bureaucratic organization is 
technically the most advanced instrument of power in the hands of 
those who control it does not as yet say anything about how force- 
fully the bureaucracy as such is capable of pushing through its con- 
ceptions within the social organization in question" (Weber, 33^, p. 
671). Bendix states even more incisively: "The bureaucracy is all- 
powerful and at the same time unable to determine how its power 
should be used" (126, p. 129). In other words, the bureaucratic 
monopoly of authority is a mere potential, mere possibility of au- 
thority without a structurally defined goal. 

Weber and Bendix have tried to explain this fact. Both of them 
argue that the indispensability of bureaucracies for the administration 
of the modern state (i.e., the very basis of their authority) has led to 
the development of a professional ethos, the dominant values of which 
are duty, service, and loyalty — in other words, values of subordina- 
tion, not of autonomous domination. Bendix emphasizes, moreover, 
the significance of formal efficiency of administration for the defi- 
nition of bureaucratic roles. Bureaucracies are oriented toward the 
smooth functioning of the machinery of administration irrespective 
of which interests are administrated according to which substantive 
principles. For bureaucracies, what exists is what exists at any given 
time; as an interchangeable constant it enters into the definition of 
bureaucratic roles. The so-called "nonpolitical civil service" is an 
expression of this attitude. Its consequences, however, are manifold 
and highly significant for conflict analysis. First, by its monopoly of 
authority unaccompanied by independent substantive interests, the 
bureaucracy of the state is, so to say, the law of inertia of social de- 
velopment become real. Bureaucracies persist as a constant force in 
all changes of the personnel of leading political positions. Following 
Weber, Bendix has emphasized the impossibility, which has its cause 
here, of revolutions in bureaucratically governed states. Secondly, 
the trend toward "purposive rationality" and "managerialism" which 
is so characteristic for the ruling values in wide areas of post-capitalist 

300 Classes in Post-Cafitalist Society (II) 

society is embodied in bureaucracies and their formalism of orienta- 
tion toward the "how," not the "what," of authority. Thirdly, the 
state administered by a bureaucracy functions almost without restric- 
tion even if political governments change rapidly, for the indispensa- 
bility of bureaucracies makes these into ever ready and ever available 
regents or temporary plenipotentiaries of authority/* 

Thus, a peculiar position must be assigned to bureaucracies in the 
light of the theory of conflict. Although they always belong to the 
ruling class, because bureaucratic roles are roles of dominance, bu- 
reaucracies as such never are the ruling class. Their latent interests 
aim at the maintenance of what exists j but what it is that exists is not 
decided by bureaucracies, but given to them. In two respects — and 
it will prove useful to separate these two aspects — the incumbents of 
bureaucratic roles are dependent on forces beyond them. First, their 
authority is borrowed or delegated authority which ultimately refers 
back to certain roles, endowed with most general authority, outside 
the orbit of bureaucracy. Although all commands are channeled, 
specified, perhaps adapted and even modified by bureaucracies, these 
commands originate outside their hierarchy. Second, the interests 
which define the substance of their authority are also given to bu- 
reaucracies. These merely administer, by virtue of their delegated 
authority, general orientations which are conceived and formulated 
elsewhere and by others. Because of this double dependence, bureauc- 
racy represents what might be called a political reserve army, a reserve 
army of authority. Domination without a bureaucracy is no longer 
possible, but domination merely by a bureaucracy is impossible, too. 
As a medium and instrument of domination, bureaucracy stands at 
the disposal of anybody who is called upon to control it. As a constant 
in political conflict it accompanies and supports whatever group is 
in power by administering its interests and directives dutifully and 
loyally. It is, as Renner rightly says, a "service class," defined by its 
service relation to authority. 

In the preceding remarks I have deliberately overstated my case 
a little. In actual fact, bureaucracies are by no means as powerless as 
our analysis might suggest. In a difi^erent context, it would be worth 
exploring the conditions of refusal of service by bureaucracies as well 

^* This is what Riesman might have meant by his remark about the state that runs 
itself (although he does not say so). Even so, a note of caution is in order. For some 
time, a state may be effectively ruled by a bureaucracy without a head; but soon the 
Impossibility of innovation will make for intense and violent conflict. This problem 
wiU be taken up again below. 

classes in Post-Cafitalist Society (II) 301 

as the structural consequences of the immanent conservatism of in- 
cumbents of bureaucratic roles. With respect to both these questions, 
Bendix has provided an approach in terms of the conditions and limits 
of bureaucratic autonomy (cf. 126, pp. 131 ff.). Many an insight 
into the regularities of social change in modern societies might be 
gained by further investigating these problems. 

Here, however, one final aspect of bureaucratic roles seems more 
important. Being incumbents of roles of dominance, the civil serv- 
ants of public administration always belong to the ruling political 
quasi-group. However, being a reserve army of authority, the bu- 
reaucratic members of this quasi-group are not potential members of 
an interest group of those in power. For structural reasons, the process 
of class formation is always abortive in the case of bureaucracies. 
Their members are, by virtue of their positions, united by certain 
(formal) latent interests. Their (substantive) manifest interests, 
however, are, again by virtue of their position, always derived from 
those of the group that represents the existing state of affairs and that 
controls bureaucracy. In principle, these manifest interests are there- 
fore interchangeable 5 they change with changing ruling groups. In 
this sense, too, bureaucracies are merely a potential j they are, in a 
paradox, a class which can never be a class, a quasi-group whose path 
to organization is blocked by the social definition of the roles of its 
members. This does not mean, of course, that civil servants cannot 
have political convictions and become members of political parties. 
But it does mean that, as civil servants — as incumbents of roles of 
dominance, that is — they cannot consistently profess a definite set of 
manifest interests. The bureaucratic reserve army of authority is a 
mercenary army of class conflict j it is always in battle, but it is forced 
to place its strength in the service of changing masters and goals. 


Who, then, constitutes the ruling class of post-capitalist society? 
The attentive reader will have noticed that we have given at least 
part of the answer already — although this simple-sounding answer 
may not meet with his approval. Obviously, we have to look for the 
ruling class in those positions that constitute the head of bureaucratic 
hierarchies, among those persons who are authorized to give directives 
to the administrative staff. This head of the hierarchy, however, is 
clearly defined: it consists of precisely those positions in the three 
branches of government which we have described above as the top 
of the pyramid of authority in the polity. The clerk in a court of law. 

302 Classes in Post-Cafitalist Society (II) 

the parliamentary private secretary, and the office director in a min- 
istry have delegated authority; they are bureaucrats. But the Speaker 
of Parliament, the Prime Minister, and the Supreme Court Judge 
have immediate authority; their decisions provide legitimation and 
substance to the work of the administrative staff. It is true that, in 
the constitutional state, judge, minister, and member of parliament 
are each responsible to somebody, too. Each is, in Laski's words, "sub- 
ject to the scrutiny of their fellow-citizens," either through repre- 
sentatives or directly through public opinion. But within the formal 
hierarchies of political authority they are the ones to whose roles 
attaches the expectation of authority, who make decisions binding upon 
all. In this sense, the leading positions of executive, legislative, and 
judiciary are the obvious places in which to look for the core of the 
ruling quasi-groups of the polity. 

I have said that this is obvious, and I should like to emphasize 
this point. It seems to me that one of the main shortcomings of 
analyses such as those by Burnham and Mills (and of Marx, for that 
matter, although there the mistake was less consequential) is that 
they do not pay enough attention to the evident seat of authority in- 
the political state, and to its occupants. Managerial or capitalist elites 
may be extremely powerful groups in society, they may even exert 
partial control over governments and parliaments, but these very facts 
underline the significance of governmental elites: whatever decisions 
are made are made either by or through them ; whatever changes are 
introduced or prevented, governmental elites are their immediate 
object or agent; whatever conflicts occur in the political arena, the 
heads of the three branches of government are the exponents of the 
status quo. It is admittedly not sufficient to identify a ruling class 
solely in terms of a governmental elite, but it is necessary to think of 
this elite in the first place, and never to lose sight of its paramount 
position in the authority structure of the state. In this sense, I should 
agree with Riesman's point about the analysis of ruling classes in post- 
capitalist society: "We cannot be satisfied with the answers given by 
Marx, Mosca, Michels, Pareto, Weber, Veblen, or Burnham, though 
we can learn from all of them" (230, p. 252). 

Of course, this insistence on governmental elites as the core of 
the ruling class must be truly shocking to anybody thinking in Marx- 
ian terms, or, more generally, in terms of the traditional concept of 
class. To some extent, the apparent incompatibility of the traditional 
and of our notion of class will be bridged presently. Basically, how- 
ever, it is important to recall at this point that the present study is 

Classes in Post-Capitalist Society (II) 303 

concerned not with the concept of class and its current applicability, 
but with the problem of conflict in industrial society. I believe that 
the traditional concept of class may be incorporated as a special case 
into the analysis suggested herej but the concept as such is irrelevant 
to our discussion. If it sounds strange that cabinet ministers, judges, 
and members of parliament should be the exponents of a ruling class, 
this strangeness is, in my opinion, due to the strangeness of reality, 
and not to the faultiness of the approach proposed in this study. 
Ruling classes, in many countries today, no longer behave as their 
predecessors did in capitalist society j they are, for this reason, no less 
ruling classes or, if one prefers these terms, dominant groups in po- 
litical conflicts. 

It is true, of course, that, much like bureaucracies, governmental 
elites provide but an incomplete description of any ruling class or 
dominant conflict group. Bureaucracies are heteronomous j they are 
therefore a reserve army of authority. Governments represent people 
and interests 5 they are therefore the exponents and not the whole of 
the ruling class. The phenomenon of representation, which compli- 
cates so many problems of political analysis in contemporary society, 
has two complementary aspects. First, governmental elites are the 
exponents of certain persons, the visible head of a body of men who 
through government have a share in authority. Cabinet members, 
judges, and members of parliament do not make up an autonomous 
group that exists by itself 5 they rule on behalf of somebody. Second, 
the substance of government policy incorporates interests which origi- 
nate outside the rather limited circle of governmental elites. The 
exercise of authority always involves both the chance to issue authori- 
tative commands and certain interests which constitute the substance 
of these commands. In order to identify those who are represented 
by governmental elites, we have to find the sources of their substantive 
policies as well as the group or groups behind them. In abstract, there- 
fore, the ruling political class of post-capitalist society consists of the 
administrative staff of the state, the governmental elites at its head, 
and those interested parties which are represented by the govern- 
mental elite. 

This conclusion is obviously so abstract and formal as to be utterly 
useless for our understanding of the societies in which we live ; and 
I do not propose to leave it at that. However, this abstract conclusion 
marks the limit beyond which a general analysis of post-capitalist 
society cannot go. All modern societies have bureaucracies and gov- 
ernments j in this respect, our conclusion holds for the United States 

304 classes in Post-Cafitalist Society (II) 

as for the Soviet Union, for West Germany as for East Germany. 
But the "interested parties" behind the governments of these coun- 
tries differ greatly. Rather than describe their composition in gen- 
eral, we must here try to determine the range of variability of 
possible structures. "Post-capitalist society" in the singular dissolves 
here into a variety of political systems. One of the most important 
of those criteria which define the range of variability of political 
conflict seems to me the homogeneity of the groups represented by 
governments. In terms of their homogeneity, we can construct a scale 
of types of ruling classes which seems to cover the empirical patterns 
found in the modern world. 

At one extreme of this scale we find the situation envisaged by 
Marx, Burnham, and others. Here, the governmental elite is part 
and parcel of a homogeneous and organized larger entity. Cabinet 
ministers, judges, and members of parliament — if such exist — are 
chosen from among this larger group, and the substance of their de- 
cisions is defined by the interests held by the larger group. Although 
not part of the formal hierarchy of political administration, the mem- 
bers of this group are incumbents of positions of political domination 
on account of their continuous, if indirect, participation in the business 
of government. This is roughly the situation described by Djilas as 
characteristic of modern totalitarian states." A state party controls 
the instruments and positions of government (including, in this case, 
the administrative staff). In terms of personnel and interests, the 
members of this party constitute the recruiting field of government. 
They are the quasi-group which attempts to maintain the status quo 
of authority relations, and their exponents at the top of the admin- 
istrative hierarchies are interchangeable representatives of identical 

At the other extreme of the scale that defines possible types of 
ruling classes we encounter the condition so aptly described by Ries- 
man. Here, the governmental elite does not represent a stable, or- 
ganized quasi-group. There are, to be sure, political parties, but these 
are themselves heterogeneous to the extent of losing their identity. 
Instead, society is split up into a large number of organizations, "veto 

^^ It has to be realized, of course, that the extremes of the scale described here 
are ideal types which we are unlikely to find in reality. If we identify, therefore, the 
two extremes with the analyses presented by Djilas (here) and Riesman (below), this 
implies a critical appraisal: both Djilas and Riesman have — perhaps deliberately, and 
in that sense justifiably — overstated their cases by claiming reality for ideal-typical 

Classes in Post-Cafitalist Society (II) 305 

groups," which "by their very nature . . . exist as defense groups, 
not as leadership groups" (230, p. 245), but which have situational 
authority. Any one act of government reveals an "issue" which has 
been advocated by somebody j veto groups enter the structure of au- 
thority when and as long as an interest or person of their choosing 
is adopted by government. This is a peculiar state of affairs, in which 
it is indeed virtually impossible to locate the ruling class. Govern- 
ment itself is here bureaucratizedj its heads merely provide channels 
of expression for varying aggregations of interests. The governing 
elites represent nobody in particular, or everybody, except in specific 
situations and decisions in which they associate with specific veto 
groups. Rather than an exponent, they are but a switchboard of 
political authority. Thus the ruling class consists of two constants, 
bureaucracy and government, and one variable, the veto group whose 
claims are, in particular situations, incorporated in government policy. 
Conversely, the subjected class consists of all those who in a given 
situation do not associate with government, but again it is a situational 
class, every element of which has its chance to exercise authority.^^ 

However, these extremes of the scale, and particularly the latter, 
are ideal types, and emphatically do not describe the conditions of any 
existing society. Within totalitarian state parties, there are invariably 
competing subgroups representing sectional interests with varying 
degrees of success. The "new class," far from being an entirely 
homogeneous entity, is a highly explosive unit in which local and 
central, industrial and agricultural, dogmatic and adaptive, bureau- 
cratic and entrepreneurial "pressure groups" struggle for domination. 
If the ruling class of present-day totalitarian societies nevertheless 
comes close to the first extreme of our scale, this is largely because 
experience shows that when the existing order itself is threatened, 
most of its subgroups abandon their differences in favor of the com- 
mon goal to uphold the status quo. 

With respect to Western societies, on the other hand, Riesman 
overstated his case in several respects. For one thing, political con- 
flict in post-capitalist society has grown out of the fierce ideological 
battles of the capitalist period, and it is at the very least unlikely that 
these have entirely ceased to play a part. For another thing, there 

^® Riesman's exposition of this situation in the chapter on "Who Has the Power? " 
(230, pp. 242-55) is really excellent, and little need be added to it, except for two 
things: first, Riesman underemphasizes the place and importance of governmental 
elites as switchboards of power j second, he describes an ideal type and not an actual 

3o6 Classes in Post-Ca'pitalist Society (II) 

still are political parties. While it is certainly true that these are less 
and less inspired by all-embracing ideologies, there still are differ- 
ences between them. Perhaps political parties, too, tend to become 
switchboards rather than holders of authority, but they are switch- 
boards with a bias. Riesman believes that "unlike a party that may 
be defeated at the polls, or a class that may be replaced by another 
class, the veto groups are always *in* " (230, p. 254). Undoubtedly, 
lobbyists of every description try to keep up good relations with all 
political parties. But for the parties, there still are hierarchies of 
interest, there is a rank order of veto groups. With respect to specific 
issues, a party cannot simultaneously satisfy trade unions and em- 
ployers' associations, veterans' organizations and pacifist groups, Prot- 
estant and Catholic churches, competing minorities, etc. Even if the 
manifest interests of political parties are reduced to an order of pre- 
cedence for lobbyists, these still provide a general direction of policy 
in matters both of personnel placement and of the substance of po- 
litical decisions. 

Thus, if the ruling class of post-capitalist countries in the West 
is "situational," this is not because it consists of ever changing veto 
groups, but because political parties "may be defeated at the polls.'' 
It has but one stable element, the bureaucracy of the state. While the 
bureaucracy is an impotent participant of political conflict in the ordi- 
nary course of affairs, its conservative effect on all modern societies, 
and especially on those whose governments change rapidly, must not 
be overlooked.^^ The governments of Western societies are often 
mere switchboards of authority j decisions are made not by them but 
through them. In this respect, the political parties from which the 
personnel of governmental elites is recruited do not differ very 
greatly from these elites. But there is, associated with every party, 
a number of veto groups that enjoy the particular favor of this party. 
If this party is in power, then the ruling class of the society in ques- 
tion consists of the four elements: bureaucracy, governmental elite, 
majority party, and its favored veto groups. If it is not in power, 
then its favored veto groups are, like its members, defense groups 
that represent the interests of the subjected class. 

^^ This eflFect was particularly evident, of course, in France before de Gaulle 
assumed power — and its consequences for political conflict are still in evidence. Gen- 
erally speaking, however, bureaucracies may promote stability, but stability and stag- 
nation are never very far apart; and where social development is arrested, political 
conflict is likely to increase rapidly in both intensity and violence. 

classes in Post-Capitalist Society (II) 307 

It might appear as if we were trying to avoid the question of who 
is the ruling class in post-capitalist society. Having first reduced its 
composition to some constant and some empirically variable elements, 
we have now described these elements in a manner exactly as formal 
as before. However, even on the reduced level of the generality of 
the last part of our analysis, we can go no further. The ruling quasi- 
groups of post-capitalist societies differ from country to country, and 
from election to election, depending on the party (and veto groups) 
in power. Otherwise put, political conflict is always situational con- 
flict between those who at a given point of time are excluded from 
authority and those who are "in." It is a characteristic feature of most 
Western societies that the "ins" and the "outs" change places, or can 
at least change places. If they do not do so for prolonged periods of 
time, if, in other words, one party stays in power for several election 
periods, it is likely that both the intensity and the violence of political 
conflict will grow, because certain interests (and veto groups) are con- 
tinuously and systematically neglected. If there is a regular exchange 
of personnel and policy, conflict will remain situational and mild, 
because all interests (and veto groups) are recognized. In any case, 
the stable elements of the ruling class of contemporary Western 
societies are politically impotent, and the dynamic elements are em- 
pirically variable. 


Many of the features characteristic of the political class structure 
of Western societies are both a consequence and an index of a process 
which, without undue extension of the term, might be described as 
political democracy. More generally, our analysis of ruling classes 
in post-capitalist societies reveals by implication different modes of 
institutionalizing political conflict j and it is to these that I propose 
to turn in the concluding sections of this study. We have seen that, 
empirically, conflict has two major aspects, those of violence and of 
intensity. These in turn determine the ways in which group conflict 
leads to structure change. How do the factors influencing violence 
and intensity of political conflict, and suddenness and radicalness of 
structure change, present themselves in post-capitalist society? Which 
future trends seem indicated in the patterns of conflict and change 
of present-day industrial countries? 

Once again — to start, in this section, with the violence aspect of 
conflict — it will help our analysis if we begin by constructing ideal 

30 8 Classes in Post-Cafitalist Society (II) 

types In order then to confront real societies with them. There is, to 
begin with, the democratic type of society. In it the conditions of 
organization are present for most subjected groups. There is freedom 
of coalition and free communication, and for most political groups 
leaders and ideologies are readily available. In fact, there are organi- 
zations — political parties — representing the interests of the opposing 
quasi-groups (with the qualifications introduced above). Moreover, 
absolute deprivation of any group of the population in terms of socio- 
economic status has become rare, if not impossible. Finally, there is 
an elaborate system of conflict regulation in the political sphere. In 
fact, the institutions of the democratic state reflect very nearly the 
model of effective conflict regulation: conflicting parties and interests 
are institutionally recognized j parliamentary bodies furnish the set- 
ting of regular conciliation between the parties j the rules of the game, 
including a constitution as well as statutory procedural arrangements, 
enable decisions to be made; certain personages, often the head of 
state, may act as mediators if autonomous conciliation breaks downj 
finally, there is the legal system to arbitrate unsolved disputes which 
threaten to break down the machinery of parliamentary negotiation. 
Under these conditions, we should expect political conflict to be 
entirely nonviolent, and structure change to be entirely gradual. 
There is always a chance for the subjected class, or its parliamentary 
representatives, to take over government, or to penetrate into the 
governing elites, or to make its claims heard and accepted without 
any change in the government personnel. The effect of political 
democracy can be convincingly illustrated by one of its crucial, and 
paradoxical, rules of the game. If, as the result of an election, the 
governing elite is replaced in its entirety and forced into the oppo- 
sition, structure change would (according to our criteria) appear to 
be extremely sudden, indeed. On a single day the personnel of all 
but the bureaucratic positions of dominance is exchanged. Why does 
this revolution fail to have the effect of a revolutionary change? A 
complete change of government in a democratic country is, to be sure, 
invariably an event of great consequence. But it is mitigated by a 
number of factors. First, there is the fact that whatever changes may 
occur, they do not affect the rules of the game agreed upon by both 
parties. Secondly, these rules of the game include for all parties the 
legitimate expectation of attaining power in the future. Finally, the 
larger part or the whole of the administrative staff provides, in all 
democratic states of the post-capitalist period, an element of stability 

classes in Post-Capitalist Society (II) 309 

that survives every change. Thus, social structure and the political 
rules of the game combine to mitigate and smooth out the comparative 
suddenness of changes resulting from elections and help to preserve 
the gradual nature of social development arising from institutional- 
ized conflict. 

This is the ideal pattern of democracy j in practice, of course, there 
are strains and distortions in it, and it would be well, here, to point 
out the various deviations from the democratic pattern — departures 
that are found in the actual political structure of all or some Western 
countries. These deviations are all the more serious since they arise 
directly from the structure of democratic institutions or from the 
rules of the game of the democratic process themselves. One example 
of such deviations has already been mentioned and discussed briefly: 
in its most limiting aspect, the process of political democracy allows 
for a rapid succession of governments which can prevent any one 
ruling group from remaining in power long enough to make its in- 
fluence felt. For all practical purposes, the top of the political hier- 
archy of authority is unoccupied. In this case, the bureaucracy of the 
state becomes its own master. Since it is unable to inaugurate anything 
new, it goes on administering the commands it was given by the last 
more permanent government. The status quo is frozen, and all parties 
and veto groups are unable to realize their interests. This is the kind 
of development which may lead an ever-growing subjected class to 
consider the abolition of "the system itself," i.e., the rules of the 
game that are (rightly, to some extent) blamed for the existing con- 

Secondly, there are no provisions in the democratic rules of the 
game to prevent a party that is supported by, say, 30 per cent or even 
45 per cent of the electorate from staying out of office indefinitely. 
As A. Downs has shown (193), this would not happen in a perfectly 
"rational" world of certainty. But in the real world there have been 
and are political parties which regularly gather a substantial and often 
stable share of the popular vote without ever becoming quite strong 
enough to form a government. This was (and still is, here and there) 
notoriously the case with Socialist parties in European countries. 
Unless such permanent minorities have alternative channels of ex- 

^® There are certain constitutional remedies, certain rules of the game, designed 
to cope with this situation; again, France (but also the German Federal Republic) 
might be cited as an example. These rules usually take the form of restrictions imposed 
on parliament with respect to voting governments out of office. 

310 Classes in Post-Ca'pitalist Society (II) 

pression," they will be increasingly alienated from "the system" that 
prevents them from getting "in" and increasingly irritated by the rules 
of the game according to which they have no right to demand a share 
in the exercise of authority. Such permanently excluded subjected 
groups are very likely to become more and more radicalized, to in- 
troduce an element of violence into political conflict, and to aim at a 
sudden replacement of all incumbents of positions of dominance — if 
not of the whole framework of political conflict that they blame for 
their ill fortunes.^" 

Finally, the system of representation in modern democratic states 
may lead to a situation in which all political conflict either is, or is 
believed to be, reduced to a struggle between competing governmental 
elites. The functionaries of political parties are so alienated from 
their members that these lose all confidence in their representatives. 
There is already, in many Western countries, a widespread feeling 
that "it does not matter for whom one casts one's vote," because "what- 
ever one votes, the same people will always rule." This state of affairs 
corresponds suspiciously closely to the dichotomous image of society 
according to which it makes no difference whether "they" call them- 
selves representatives of the workers or of the employers. It also 
corresponds to the actual collusion which is so general a feature among 
the representatives of political parties. What happens, here, is that 
the ruling class becomes a small elite of functionaries of nominally 
different or even conflicting organizations, while the overwhelming 
majority of the people form a subjected class whose every access to 
authority is blocked. The blocking is done by their so-called repre- 
sentatives who, when the position of the "insiders" is threatened with 
dislodgment, jealously guard the "inside" position not only of them- 
selves but even of their competitors. Under these conditions, the sub- 
jected class has no channel of organization but one: the one described 
by Marx in his "i8th Brumaire of Louis Bonaparte." There may 
come to the fore one man, or a small group of men, who succeeds in 
convincing the many that they are the "real" representatives of their 

^® Such as a federal political structure which enables them to control at least state 
governments. This possibility serves as a compensating force, e.g., in West Germany 
and in Canada (cf. Lipset, 1 99). 

^° There are no formal rules of the game to provide for this eventuality; but there 
is the possibility of the (presumably equally permanent) government incorporating 
demands of the permanent opposition to its policies so as to thaw the fronts of conflict 
somewhat. In any case, a permanent minority — Socialist, Communist, or otherwise — 
remains a disturbing element of political democracy. 

classes in Post-Captalist Society (II) 311 

interests, and who promises to do away with the whole "system" of 
parties, parliaments, and democratic governments. Whatever the sub- 
stantive outcome of such a process, it is certain that it will lead to 
increasingly violent conflicts.^^ 

It would seem that all the inherent dangers of political democracy 
point to an alternative of conflict regulation which may be described 
as an ideal type opposite to democracy. In the terms of classical po- 
litical thought this other extreme of the scale is oligarchy, or tyranny. 
In a modern state organized along oligarchic and authoritarian lines, 
the conditions of organization for opposing groups are absent. There 
is but one "official" organization, if any, and outside this organization 
freedom of coalition is drastically curtailed. Above all, however, such 
states have no accepted routines of conflict regulation, because their 
governments do not recognize the legitimacy of political conflict itself. 
Wherever there are signs of "rebellion" these are forcefully sup- 
pressed. The rule of the few is upheld at all costs. 

Conflict theory would suggest that this kind of political organi- 
zation would make for extremely violent political conflicts as well as 
sudden structure changes. However, this is clearly a self-contradict- 
ing conclusion. There cannot be violent conflict and eff^ective sup- 
pression at the same time. Either one or the other, the principles of 
either conflict theory or of the ideal type of tyranny, must be wrong 
if their results contradict each other. I suggest that what is wrong 
here is the picture of the ideal type: there is, in reality, no country 
in which all conflict is eiSFectively suppressed by force, and in which 
the arbitrary rule of a small group is upheld by sheer brutality. In 
fact, I should claim that one of the important consequences of the 
approach to conflict analysis proposed in this study is that it exposes 
as mistaken the view that there are political systems based on force 

^^ Considering Michels' "iron law of oligarchy," the latter perspective would 
seem to be not only a very real, but almost an inevitable, result of representative democ- 
racy. "If we accept Michels' formulation of the problem, we are faced with the fact 
that modern democracies do not constitute a government by the people. Instead, this 
form of government consists of conflicts of power among oligarchic organizations, with 
the result that the average citizen exercises his rights of citizenship only in the sense 
that he chooses among alternatives presented to him by these competing oligarchies" 
(Bendix and Lipset, 191, p. 95). This would, however, be an overdrawn conclusion. 
Depending on the rules of the game in force, some countries are in fact more affected 
than others by the tendency toward collusion among the representatives: those with 
proportional representation more than those with a relative majority system of elec- 
tion; those with a multiparty system more than those with a two-party system. Once 
again, however, there is no absolute safeguard. 

312 Classes in Post-Cafitalist Society (II) 

alone. Instead, conflict theory directs our attention to those phe- 
nomena in existing nondemocratic countries which can only be under- 
stood as attempts to regulate political conflict. More than in the case 
of democratic countries, such approximations to the ideal type of 
tyranny as there are deviate from its patterns. 

It is obvious that in those countries (post-capitalist or otherwise) 
which are under Communist rule, the conditions of organization are 
lacking, there is to some extent absolute deprivation, and political 
conflict is not officially recognized as a social force. It is therefore 
also true that a tendency toward violent conflict and sudden change 
accompanies these totalitarian countries in every phase of their de- 
velopment. After the events of the last decade, there is no need to 
elaborate on these facts — which, by implication, testify to the ulti- 
mate ineffectiveness of brutal suppression. However, the modern 
tyranny called totalitarianism is not as simple as we have made it here. 
Suppression and nonrecognition of conflict are merely the visible facets 
of a far more complex reality. 

In our attempt to identify the dominant groups of contemporary 
societies, we have described the ruling class of totalitarian countries 
as particularly homogeneous and organized. By virtue of this fact 
there is, in totalitarian countries, a clear line of demarcation between 
the rulers and the ruled. Where the party ends, the subjected class 
begins. And the subjected class constitutes a large quasi-group whose 
many partially divergent interests are combined into one demand — a 
change in the status quo of authority. But this group is unable to 
organize itself; there is no freedom of coalition. Whatever conflicts 
exist are therefore forced to remain latent or under cover. Yet even 
the "official" structure of political authority provides this subjected 
quasi-group with channels of expression. In order to recognize these, 
we must remember that "discussion" is one of the crucial features of 
totalitarian government. Nowhere is there as much "discussion" as 
in the one-party countries of the modern world: meetings in one's 
factory or office, street or house, trade union, cooperative society, choir 
or football club, school, etc., serve the one purpose of "discussing" 
things. These "discussions" are not, to be sure, opportunities for a 
free exchange of ideas. They are above all attempts at indoctrination 
and at soliciting that brand of "voluntary cooperation" so peculiar to 
modern totalitarian states. But inter alia and in a minor way the 
meetings and "discussions" which loom so large in the life of every 
subject of totalitarian government provide a chance to voice, cautiously 
and in the accepted language, criticisms of individuals and policies, 

Classes in Post-Cafitalist Society (II) 313 

suggestions, and demands. From the point of view of the ruling 
class, this fact is both "functional" and "dysfunctional." On the one 
hand, the party organization and its varied affiliations serve as a 
gigantic institute of opinion research which, through meetings and 
"discussions," tries to explore the "wishes and feelings of the people." 
On the other hand, the same meetings and "discussions" that are 
necessary for this as well as for many another purpose bring into 
contact the otherwise scattered members of the subjected quasi-group 
and form the nuclei of actual and future conflicts. It is no accident 
that the revolts in Communist countries originated among those who 
meet most often in large numbers: building workers, steel workers, 

Undoubtedly, meetings and "discussions" as channels of conflict 
are, from the point of view of the ruling class, largely a pretext. It 
may well be that state parties explore their subjects' wishes and feel- 
ings not in order to incorporate these in their policies, but in order to 
see how much further they can push their own plans. However, here 
again the ruthlessness of oligarchic rulers must not be exaggerated. 
The skillfully ruled totalitarian countries"' are by no means stagnant. 
Change and development play almost as great a part in them as they 
do in democratic societies. Here, the institution of "purges" is crucial. 
Some naive observers believe that purges within the new class of to- 
talitarian countries are merely an expression of personal or group 
rivalries. These may in fact provide the motivation of those who are 
either actively or passively engaged in them, but purges serve another 
purpose as well. Like elections in democratic states they involve a 
partial replacement of the governing elite, and this replacement means 
in fact, or appears to mean to the subjected class, a change of policy 
which incorporates interests hitherto unacknowledged. We have de- 
fined structure change operationally in terms of changes affecting the 
personnel of leading authority positions. I would insist on this defini- 
tion, and I suggest at the same time that an investigation of frequen- 
cies and rates of change in democratic and totalitarian countries would 
reveal very little difference between the two. 

All this is emphatically not to say that totalitarian political systems 
are not as bad as they are made out to be in the West. It is to say, 

^^ Apart from using this mechanism of channeling social conflict, totalitarian 
rulers are of course experts in redirecting conflict energies and providing safety-valve 
institutions. Coser's study of conflict (8l) supplies many examples of such patterns. 

^^ Nazi Germany was not, in this sense, "skillfully ruled"; but most Communist 
countries of eastern Europe are. 

314 classes in Post-Capitalist Society (II) 

however, that we must appraise their political structure realistically, 
and that the sociological theory of conflict can help us do so. While 
there is an undercurrent of potentially violent conflict aiming at sud- 
den change in all totalitarian countries, it does not follow that this 
latent antagonism will break out into the open. The reason for this is 
not, as so many believe, the eflFectiveness of suppression in totalitarian 
states, but the existence of a hidden system of conflict regulation in 
which meetings and "discussions" as well as purges play an important 
part. Horrible as this idea may be, it is possible to establish a totali- 
tarian rule in which the violence of political conflict is reduced almost 
as effectively as by the democratic process of conflict regulation.^* 
Suppression of conflict, however, far from helping to control politi- 
cal conflict, defeats its own ends: totalitarian governments are in 
"danger" of being violently overthrown to the extent to which they 
resort to suppression as a means of dealing with conflict. 


Their modes of conflict regulation constitute one of the most im- 
portant distinguishing features of democratic and totalitarian political 
systems. By looking at these in terms of the institutional arrangements 
set up in order to control political conflict, we are able to relate dif- 
ferent forms of government to different patterns of social structure. 
This is a fortiori the case with respect to the factors that influence the 
intensity of political conflict. The antithesis of pluralism vs. monism 
with respect to associations, scales of differentiation, and patterns of 
conflict seems especially suited to describe the social bases of politi- 
cal freedom. A free society encourages diversity in its institutions 
and groupings to the extent of actually promoting divergence j con- 
flict is the life breath of freedom. A totalitarian society insists on 
unity to the extent of uniformity j conflict is a threat to its coherence 
and survival. 

First, the intensity of political conflict depends on the presence 
of the conditions of organization. We have already seen that while in 
free societies subjected quasi-groups are free to organize themselves 

^* The "almost" in this statement is of course deliberate. Clearly, the danger of 
systematically ignoring certain interests and groups is much greater for even the 
cleverest totalitarian government than it is for the government of a country in which 
everybody is free to voice his discontent. Here, as elsewhere, it is safer to trust the 
spontaneity of individuals and groups than to trust a "plan." In any case, Communist 
countries have not succeeded in setting up an eflfective hidden machinery of conflict 

Classes in Post-Capitalist Society (II) 315 

in defense of their interests, their counterparts in totalitarian societies 
have to search out devious channels to make possible even a minimum 
of expression of their interests. Due to the lack of the conditions of 
organization, there is a permanent, and often growing, quantity of 
unreleased pressure in totalitarian states which imbues their latent 
political conflicts with an intensity unknown in free societies, where 
pressure is released almost as soon as it is created. There can be little 
doubt that repression of conflict by force raises the cost of victory or 

A second determinant of the intensity of conflict is the amount of 
mobility found in a given society. Without once again entering into a 
detailed discussion of available evidence, we can maintain that, in the 
post-capitalist societies of the West, rates of social mobility are fairly 
high. There are still, to be sure, barriers in the system of social strati- 
fication which are hard for the individual to overcome j but the educa- 
tional institutions operate as agents of role allocation, and such biases 
as there are in the educational system are increasingly being removed. 
Mobility through education includes mobility between the classes: 
from manual to clerical and professional occupations, and vice versa. 
Insofar as intergeneration mobility constitutes the predominant type 
of fluctuation, class position may still be a largely inescapable reality 
for the individual 5 but it is no longer a collective fate. The hope that 
one's son might rise to the position one has oneself failed to attain is 
one of the most widespread feelings of people in post-capitalist so- 
ciety. Insofar as mobility affects it, the intensity of political conflict 
has diminished considerably in the last decades. 

The extent of social mobility in totalitarian societies is not alto- 
gether easy to assess. For all we know, there is a high amount of move- 
ment up and down the status scale in existing totalitarian countries; in 
the East, as in the West, the educational system plays an important 
part. Some observers of Soviet society have claimed that in Russia 
social mobility is indeed so widespread as to make for the type of 
"classlessness" described above as the society made classless by mo- 
bility. This, however, is open to doubt. T. Bottomore, who reports 
a statement to this eflFect, holds two convincing objections against it. 
First, "the high level of mobility may ... be regarded as a phe- 
nomenon resulting from the needs of industrial development and not 
any conscious attempt to promote it" (37, p. 42). Thus, many of the 
present-day totalitarian countries are in fact not post-capitalist socie- 
ties but societies at an early stage of industrialization, and they there- 
fore display the symptoms characteristic of the industrial revolution. 

3 1 6 Classes in Post-Cafitalist Society (II) 

Secondly, "there is in fact some evidence that social mobility is now 
being restricted" (37, p. 42). In any case, "there is no evidence that 
social mobility is, at the present time, any greater than in the Western 
democracies, and much evidence that it is being deliberately reduced" 
(p. 43). If we can believe Djilas, this reduction of mobility would 
seem, moreover, to involve, above all, the "new class" in totalitarian 
countries which increasingly closes itself to any influx from other 
strata, especially from the subjected class. We can assume that to the 
extent that this trend of closure continues, political conflict in totali- 
tarian countries is becoming more intense. 

Of all factors influencing the intensity of political conflict, those 
of the pluralism-superimposition scale are by far the most efFective. 
In order to assess their consequences, I shall resort, for a last time, to 
constructing contrasting ideal types. At one extreme of the scale thus 
emerging, we should find a society in which all patterns, issues, and 
contexts of political conflict are superimposed and combined into two 
large hostile camps. There is superimposition with respect to the 
structure of authority and to the scales of rewards that make up social 
stratification. Whoever occupies a position of authority has wealth, 
prestige, and other emoluments of social status at his disposal, tooj 
whoever is excluded from political authority has no hope of climbing 
very far on the scale of social status. Furthermore, the conflicts aris- 
ing from different associations are superimposed. Power is generalized 
in the sense that a homogeneous and interchangeable elite governs an 
identical subjected class in the state, in industry, in the army, and in all 
other associations. Finally, such nonclass conflicts as exist in society 
are congruent with the conflicts arising out of the unequal distribu- 
tion of authority. Political class conflict, industrial class conflict, re- 
gional conflicts, conflicts between town and country, possibly racial and 
religious conflicts — all are superimposed so as to form a single and 
all-embracing antagonism. Under these conditions, the intensity of 
political conflict reaches its maximum. 

As with our earlier ideal types, there is no actual society in which 
this is fully realized. There are, however, many indications that 
superimposition of patterns, and the monism of social structure re- 
sulting from it, are a characteristic feature of modern totalitarian 
states. The ruling and the subjected groups of industry and the state 
are identical j the party exercises its power in both associations. With 
respect to the military, the same condition is aimed at, although — as 
the Russian example shows — it is not attained without a struggle. In 
any case, generalized authority is the constant goal of the state party. 

classes in Post-Capitalist Society (II) 317 

This generalized authority includes the attempt to monopolize every 
scale of socioeconomic status for the "new class." Its members enjoy, 
apart from, or perhaps by virtue of, their authority, high incomes and 
considerable prestige — although the latter is less subject to manipula- 
tion than the former.^" With respect to the superimposition of class 
and other conflicts, no unequivocal pattern seems discernible in to- 
talitarian countries, although a tendency to alienate minorities as well 
as subjected groups in terms of authority is not uncommon under 
oligarchic rule. Divide et im'pera is an old and supposedly useful im- 
perative of despotism} but with respect to the social structure of their 
countries, present-day totalitarian rulers have not followed its pre- 
scription. They have, instead, aimed at a uniform and monistic or- 
ganization of society. Moreover, in the Communist countries of the 
East they have adopted an ideology of intense conflict. The two go 
together well: on account of the monistic structure of conflict (and 
many other) relations in totalitarian countries, these relations have 
gained, and continue to gain, in intensity. Whatever conflicts do oc- 
cur involve both rulers and ruled with their whole personalities j and 
if these conflicts become open and violent, the cost of defeat is too 
high for both parties to allow graceful retreat. "Totality" distin- 
guishes the totalitarian state in more than one respect, including the 
extent of the changes desired by those who, for the time being, are its 
powerless subjects. 

The ideal type opposite to that of totalitarianism is that of a free 
society. Here, the intensity of political conflict is reduced to a mini- 
mum. The scales of social stratification are largely separate } posses- 
sion of authority does not necessarily imply wealth, prestige, security. 
There are competing elites at the top of the various scales. Conflicts 
in different associations are dissociated. Leadership in the state does 
not imply leadership in industry, in the army, or in other associations, 
nor does exclusion from authority in one context imply exclusion in 
all others. Class conflict and other clashes between groups are dis- 
sociated, too; being a member of a particular minority, race, or church 
does not automatically convey certain privileges or disabilities with 
respect to the distribution of political authority. Pluralism of institu- 
tions, conflict patterns, groupings, and interests makes for a lively, 
colorful, and creative scene of political conflict which provides an op- 
portunity for success for every interest that is voiced. 

^■^ They do not enjoy high "job security" nor will they be able to achieve this 
under totalitarian conditions. In view of possible revolts and the role of the ruling 
class in them, this fact must not be underestimated. 

3 1 8 Classes in Post-Ca-pitalist Society (II) 

Needless to say, there is no society which corresponds in all re- 
spects with this ideal-typical picture. With respect to minorities, In 
particular, there still is a great deal of superimposition of conflict 
fronts in all Western countries. Being a member of this or that church, 
race, or ethnic group puts many people at a disadvantage in the strug- 
gle for political authority. Regarding the other two factors, how- 
ever, we have seen that a pluralist structure of society is in fact pro- 
gressing in the post-capitalist world. One of its symptoms consists in 
the institutional isolation of industry j this in turn involves some dis- 
sociation of the scales of wealth and authority. By these and similar 
factors, the involvement of people in political conflict decreases j in- 
dividuals, veto groups, and political parties can, so to speak, afford to 
lose 5 and if they win, the changes they introduce are piecemeal rather 
than radical. History is a permanent guest in a free society, not an 
unwanted intruder whose presence signals revolutionary upheavals. 

I concur with the common belief that the struggle between free 
societies and totalitarian societies is the dominant issue of political con- 
flict in our time. Contrary to many, however, I do not believe that 
this struggle is confined to international relations. The struggle be- 
tween freedom and totalitarianism occurs within societies as well as 
between them. No real society can be found at one or the other ex- 
treme of our ideal types. There are very nearly free and very nearly 
totalitarian countries, but more often we find intermediate forms. In 
the modern world, there are such paradoxical states as democracy 
without liberty and liberty without democracy. Everywhere, how- 
ever, the struggle between freedom and totalitarianism may be re- 
garded as one between different attitudes toward social conflict. To- 
talitarian monism is founded on the idea that conflict can and should 
be eliminated, that a homogeneous and uniform social and political 
order is the desirable state of affairs. This idea is no less dangerous 
for the fact that it is mistaken in its sociological premises. The plural- 
ism of free societies, on the other hand, is based on recognition and 
acceptance of social conflict. In a free society, conflict may have lost 
much of its intensity and violence, but it is still there, and it is there 
to stay. For freedom in society means, above all, that we recognize 
the justice and the creativity of diversity, difference, and conflict. 




1. Friedrich Engels, Die Lage der arbeitenden Klassen in England. New ed. 

Berlin, 1952. 

2. Adam Ferguson, Essay on the History of Civil Society. London, 1767. 

3. Karl Marx, Nationalökonomie und Philosophie. Ed. by E. Thier. Cologne 

and Berlin, 1950. 

4. Die heilige Familie, oder Kritik der kritischen Kritik. In Marx- 
Engels Gesamtausgabe, Section I, Vol. III. Berlin, 1932. 

5. "Die moralisierende Kritik und die kritische Moral," in Franz Mehr- 

ing, ed., Aus dem literarischen Nachlass von Karl Marx und Friedrich 
Engels. 3d ed. Stuttgart, 1920. 

6. Das Elend der Philoso f hie. New ed. Berlin, 1947. 

7. Zur Kritik der 'politischen Ökonomie. New ed. Berlin, 1947. 

8. Der 18. Brumaire des Louis Bonaparte. New ed. Berlin, 1946. 

9. Die Klassenkämpfe in Frankreich. New ed. Berlin, 1951. 

10. Letter to Bolte (November 1871), in Zur Kritik des Gothaer Pro- 
gramms. New ed. Berlin, 1946. 

1 1. "Zirkularbrief" to Bebel and others (September 1879), in Zur Kritik 

des Gothaer Programms. New ed. Berlin, 1946. 

12. Das Kapital. New ed. Berlin, 1953. 

13. Karl Marx and Friedrich Engels, "Die deutsche Ideologie," in Siegfried 

Landshut, ed.. Der historische Materialismus. Stuttgart, 1953. 

14. Manifest der kommunistischen Partei. New ed. Berlin, 1953. 

15. John M-iHzi, An Historical Viezv of the English Government. London, 1787. 

16. Thorstein Veblen, The Theory of the Leisure Class. New York, 1899. 


17. Gerhard Albrecht, Die sozialen Klassen. Leipzig, 1926. 

18. Käthe Bauer-Mendelberg, "Stand und Klasse," Koelner Vierteljahreshefte für 

Soziologie, Vol. III, No. 4 (1924). 

19. Pontus Fahlbeck, Die Klassen und die Gesellschaft. Jena, 1922. 

20. Theodor Geiger, "Zur Theorie des Klassenbegriffs und der proletarischen 

Klasse," in Schtnollers Jahrbuch, Vol. LIV (1930). 

21. Georg hukics, Geschichte und Klassenbetvusstsein. Berlin, 1923. 

22. Robert Michels, "Beitrag zur Lehre von der Klassenbildung," Archiv für 

Sozialzvissenschaft, Vol. XLIX (1920). 

320 Bibliography 

23. Paul Mombert, "Zum Wesen der sozialen Klasse," in Hauftfrobleme der 

Soziologie: Erinnerungsgabe für Max Weber. Munich, 1923. 

24. Gaetano Mosca, Die herrschende Klasse. Bern, 1950. 

25. Vilfredo Pare to, Allgemeine Soziologie. Transl. and ed. by C. Brinkmann. 

Tübingen, 1955. 

26. Karl Renner, Die Wirtschaft als Gesamtfrozess und die Sozialisierung. Ber- 

lin, 1924. 

27. J. A. Schumpeter, "Die sozialen Klassen im ethnisch homogenen Milieu" 

(igzj), m Aufsätze zur Soziologie. Tübingen, 1953. 

28. Werner Sombart, Der moderne Kapitalismus. 4th ed. Munich and Leipzig, 


29. "Die Idee des Klassenkampfes," Schriften des Vereins für Sozial^ 

folHik, Vol. CLXX (1925). 

30. Pitirim Sorokin, Contemporary Sociological Theories. New York, 1928. 

3 1 . Othmar Spann, "Klasse und Stand," in Handwörterbuch der Staatswissen- 

schaften, Vol. V. 4th ed. 1923. 

32. Max Weber, "Der Sozialismus," in Gesammelte Aufsätze zur Soziologie und 

Sozialpolitik. Tübingen, 1924. 
330. The Theory of Social and Economic Organization. Transl. by A. M. 

Henderson and Talcott Parsons. New York, 1950. 
333. Wirtschaft und Gesellschaft (Grundriss der Sozialökonomik, section 

III). 4th ed. Tübingen, 1947. 


34. Raymond Aron, "Social Structure and the Ruling Class," in 35 below. 

35. Reinhard Bendix and S. M. Lipset, eds., Classy Status and Power: A Reader 

in Social Stratification. Glencoe, 1953. 

36. Reinhard Bendix and S. M. Lipset, "Karl Marx's Theory of Social Classes," 

in 35 above. 

37. T. B. Bottomore, Classes in Modern Society. London, 1955. 

38. Richard Centers, The Psychology of Social Classes. Princeton, 1949. 

39. G. D. H. Cole, Studies in Class Structure. London, 1955. 

40. O. C. Cox, "Estates, Social Classes and Political Classes," American Sociologi- 

cal Review, Vol. X (1945). 

41. Ralf Dahrendorf, "Gibt es noch Klassen? Die Begriffe der 'sozialen Schicht' 

und 'sozialen Klasse' in der Sozialanalyse der Gegenwart," Annales Uni- 
versitatis Saraviensis, section Philosophie, Vol. II, No. 4 (1954). 

42. "Klassenstruktur und Klassenkonflikt in der entwickelten Industrie- 
gesellschaft," Die neue Gesellschaft, Vol. II, No. 4 (1955). 

43. "Social Structure, Class Interests and Social Conflict," in Transac- 
tions of the Third World Congress of Sociology, Vol. III. London, 1956. 

44. Milovan Djilas, The New Class: An Analysis of the Communist System. New 

York, 1957. 

45. P. N. Fedoseyev, "Laws of Social Changes in the Twentieth Century," in 

Transactions of the Third World Congress of Sociology, Vol. VIII. Lon- 
don, 1957. 

46. Theodor Geiger, Die Klassengesellschaft im Schmelztiegel. Cologne and 

Hagen, 1949. 

Bibliography 321 

47. Morris Ginsberg, Sociology. London, 1953. 

48. M. M. Gordon, "Social Class in American Society," American Journal of 

Sociology^ Vol. LV (1950). 

49. Llewellyn Gross, "The Use of Class Concepts in Sociological Research," Amer- 

ican Journal of Sociology, Vol. LIV (1949), 

50. Georges Gurvitch, "Le Dynamisme des classes sociales," in Transactions of the 

Third World Congress of Sociology, Vol. III. London, 1956. 

51. yiAVirxce YizVowzchi, Les Classes sociales. Paris, 1937. 

52. Richard Hoggart, The Uses of Literacy. London, 1957. 

53. Siegfried Landshut, "Die Gegenwart im Lichte der Marxschen Lehre," in 

H.-D. Ordieb, ed., Hamburger Jahrbuch für Wirtschafts- und Gesell- 
schaf tsfolitik, Vol. I .Tübingen, 1956. 

54. G. E. Lenski, "American Social Classes — Statistical Strata or Social Groups?" 

American Journal of Sociology, Vol. LVIII (1952). 

55. S. M. Lipset and Reinhard Bendix, "Social Status and Social Structure," British 

Journal of Sociology, Vol. II (1951). 

56. T. H. Marshall, ed.. Class Conflict and Social Stratification. London, 1 93 8. 

57. Citizenship and Social Class. Cambridge, 1950. 

58. "Social Class: A Preliminary Analysis," in 57 above. 

59. "The Nature of Class Conflict," in 56 above. (See also 35 and 57.) 

60. K. B. Mayer, "The Theory of Social Classes," in Transactions of the Second 

World Congress of Sociology, Vol. II. London, 1954. 

61. Class and Society. New York, 1955. 

62. C. W. Mills, The New Men of Power. New York, 1954. 

63. The Power Elite. New York, 1956. 

64. V. S. Nemchinov, Changes in the Class Structure of the Population of the 

Soviet Union. Mimeographed for the Third World Congress of Sociology 

65. "Changes in the Class Structure of the Population of the Soviet 

Union" (abridged version), in Transactions of the Third World Congress 
of Sociology, Vol. Vlll. London, 1957. 

66. Stanislaw Ossowski, "Old Notions and New Problems: Interpretations of Social 

Structure in Modern Society," in Transactions of the Third World Con- 
gress of Sociology, Vol. III. London, 1956. 

67. Talcott Parsons, "Social Classes and Class Conflict in the Light of Recent Socio- 

logical Theory," in Essays in Sociological Theory. Rev. ed. Glencoe, 

68. Heinrich Popitz, "Zum Begrifl!" der Klassengesellschaft," in H.-D. Ortlieb, ed., 

Hamburger Jahrbuch für Wirtschafts- und Gesellschaftspolitik, Vol. III. 
Tübingen, 1958. 

69. Heinrich Popitz, H. P. Bahrdt, E. A. Jueres, and H. Resting, Das Gesell- 

schaftsbild des Arbeiters. Tübingen, 1957. 

70. Karl Renner, Mensch und Gesellschaft: Grundriss einer Soziologie. Vienna, 


71- Wandlungen der modernen Gesellschaft: zwei Abhandlungen über 

die Probleme der Nachkriegszeit. Vienna, 1953. 

72. Hehnut Schelsky, "Gesellschaf d icher Wandel," Offene Welt, No. 41 (1956). 

73. J. A. Schumpeter, Cafitalism, Socialism and Democracy. London, 1943. 

322 Bibliography 

74. Paul Sering, Jenseits des Kafitalismus. Nürnberg, 1 947. 

75. W. L. Warner, Social Class in America. Chicago, 1949. 

76. Alfred Willener, Images de la societe et classes sociales. Bern, 1 95 7. 


77. "Approaches to the Study of Social Conflict: A Colloquium," Confiict Reso- 

lution, Vol. I, No. 2 (June 1957). 

78. John Bowlby, "A Psychoanalytic View of Conflict and Its Regulation" 

(mimeo.). Center for Advanced Study, 1958. 

79. J. S. Coleman, Comtnunity Conflict. Glencoe, 1957. 

80. L. A. Coser, "Social Conflict and Social Change," British Journal of Sociology, 

Vol. VII, No. 3 (September 1957). 

81. The Functions of Social Conflict. London, 1 956. 

82. Ralf Dahrendorf, "Toward a Theory of Social Conflict," Conflict Resolution, 

Vol. II, No. 2 (June 1958). 

83. International Sociological Association, The Nature of Conflict. With contri- 

butions by J. Bernard, T. H. Pear, R. Aron, and R. C. Angell. Paris, 
UNESCO, 1957. 

84. Clark Kerr, "Industrial Conflict and Its Mediation," American Journal of 

Sociology, Vol. LX, No. 3 (November 1954). 

85. A. W. Kornhauser and others, eds., Industrial Conflict. New York, 1954. 

86. David Lockwood, "Arbitration and Industrial Conflict," British Journal of 

Sociology, Vol. VI, No. 4 (1955). 

87. H. L. Sheppard, "Approaches to Conflict in American Industrial Sociology," 

British Journal of Sociology, Vol. V, No. 4 (1954). 



88. C. J. Barnard, "The Functions and Pathology of Status Systems in Formal 

Organizations," in W. F. Whyte, ed.. Industry and Society. New York 
and London, 1946. 

89. A. M. Carr-Saunders and D. C. Jones, A Survey of the Social Structure of Eng- 

land and Wales. 2d ed. Oxford, 1937. 

90. Kingsley Davis and W. E. Moore, "Some Principles of Stratification," in 

Logan Wilson and W. L. Kolb, Sociological Analysis. New York, 1949. 

91. Theodor Ge'igcr, Die soziale Schichtung des deutschen Volkes. Stuttgart, 1932. 

92. D. G. MacRae, "Social Stratification (Bibliography)," Current Sociology, 

Vol. II (1953/54). 

93. Heinz Markmann, "Sozialstruktur im Allgemeinen," in Alfred Weber, ed., 

Einführung in die Soziologie. Munich, 1955. 

94. T. H. Marshall, "A Note on 'Status,' " in K. M. Kapadia, ed., Professor 

Ghurye Felicitation Volume. Bombay, 1954. 

95. "General Survey of Changes in Social Stratification in the Twen- 
tieth Century," in Transactions of the Third World Congress of Sociol- 
ogy, Vol. III. London, 1956. 

96. Stanislaw Ossowski, "La Vision dichotomique de la stratification sociale," 

Cahiers Internationaux de sociologie, Vol. XX (1956). 

Bibliography 323 

97. H. W. Pfautz, "The Current Literature on Social Stratification: Critique and 

Bibliography," American Journal of Sociology, Vol. LVIII (1952). 

98. Helmut Schelsky, "Die Bedeutung des Schichtungsbegriffes für die Analyse 

der gegenwärtigen deutschen Gesellschaft," in Transactions of the Sec- 
ond World Congress of Sociology, Vol. I. London, 1954. 

99. J. A. Schumpeter, "Das soziale Antlitz des Deutschen Reiches" (1929), in 

Aufsätze zur Soziologie. Tübingen, 1953. 

100. W. L. Warner and P. S. Lunt, The Social Life in a Modern Community. 

New Haven, 1941. 


101. Reinhard Bendix and S. M. Lipset, Social Mobility in Industrial Society. 

Berkeley and London, 1958. 

102. K. M. Bolte, "Ein Beitrag zur Problematik der sozialen Mobilität," Kölner 

Zeitschrift für Soziologie und Sozialfsychologie, Vol. VllI, No. l (1956). 

103. Sozialer Aufstieg und Abstieg. Eine Untersuchung über Berufs- 

f restige und Berufsmobilität. Stuttgart, 1959. 

104. "Some Aspects of Social Mobility in Western Germany," in Trans- 
actions of the Third World Congress of Sociology, Vol. III. London, 1956. 

105. Ralf Dahrendorf, "Die soziale Funktion der Erziehung in der industriellen 

Gesellschaft," Sfeculum: Saarländische Studentenzeitschrift, Vol. I, No. 7 

106. Jean E. Floud, "The Educational Experience of the Adult Population of Eng- 

land and Wales as at July 1949," in 1 07 below. 

107. D. V. Glass, ed., Social Mobility in Britain. London, 1 954. 

108. D. V. Glass and J. R. Hall, "Social Mobility in Britain: A Study of Inter- 

Generation Changes in Status," in 107 above. 

109. J. R. Hall and D. C. Jones, "The Social Grading of Occupations," British 

Journal of Sociology, Vol. I (1950). 

1 10. J. R. Hall and D. V. Glass, "Education and Social Mobility," in 107 above. 

111. J. R. HaU and W. Ziegel, "A Comparison of Social Mobility Data for England 

and Wales, Italy, France, and the U.S.A.," in 107 above. 

112. Hilde T. Himmelweit, "Social Status and Secondary Education since the 1944 

Act: Some Data for London," in 107 above. 

113. Alex Inkeles and P. H. Rossi, "National Comparisons of Occupational Pres- 

tige," American Journal of Sociology, Vol. LXII, No. 4 (1956). 

1 1 4. Morris Janowitz, "Some Consequences of Social Mobility in the United States," 

Transactions of the Third World Congress of Sociology, Vol. III. London, 

115. Japan Sociological Society, Social Mobility in Jafan. Mimeographed for the 

Third World Congress of Sociology (1956). 

116. S. M. Lipset and H. L. Zetterberg, "A Theory of Social Mobility," in Trans- 

actions of the Third World Congress of Sociology, Vol. III. London, 

117. Gerhard Mackenroth and K. M. Bolte, "Bericht über das Forschungsvorhaben 

'Wandlungen der deutschen Sozialstruktur (am Beispiel Schleswig-Hol- 
stein),' " in Transactions of the Second World Congress of Sociology, 
Vol. II. London, 1954. 

324 Bibliografhy 

118. C. A. Moser and J. R. Hall, "The Social Grading of Occupations," in 107 


119. Ramakrishna Mukherjee, "A Study of Social Mobility between Three Genera- 

tions," in 107 above. 

120. National Opinion Research Center, "Jobs and Occupations: A Popular Evalua- 

tion," in 35 above. 

121. Natalie Rogoff, Recent Trends in Occupational Mobility. Glencoe, 1953. 

122. Helmut Schelsky, Soziologische Bemerkungen zur Rolle der Schule in unserer 

Gesellschaftsverfassung. Unpublished, mimeographed (1956). 

123. G. Thomas, "Labour Mobility in Great Britain 1945—49" (An Inquiry car- 

ried out for the Ministry of Labour and National Service), Refort No. 
134 of The Social Survey. London, n.d. 


124. H. p. Bahrdt, Industriebürokratie: Versuch einer Soziologie des industri- 

alisierten Bürobetriebes und seiner Angestellten. Stuttgart, 1958. 

125. Joseph Ben-David, "The Rise of a Salaried Professional Class in Israel," in 

Transactions of the Third World Congress of Sociology, Vol. III. London, 

126. Reinhard Bendix, "Bureaucracy and the Problem of Power," in 136 below. 

127. John Bonham, The Middle Class Vote. London, 1954. 

128. Lewis Corey, The Crisis of the Middle Class. New York, 1935. 

129. Fritz Croner, Die Angestellten in der modernen Gesellschaft. Frankfurt a.M. 

and Vienna, 1954. 

130. Michel Crozier, "Le Role des employes et des petits fonctionnaires dans la 

structure sociale frangaise contemporaine," in Transactions of the Third 
World Congress of Sociology, Vol. III. London, 1956. 

131. Alvin Gouldner, "Eine Untersuchung über administrative Rollen," in Kölner 

Zeitschrift für Soziologie und Sozialfsychologie, Vol. VIII, No. l (1956). 

132. R. K. Kelsall, D. Lockwood, and A. Tropp, "The New Middle Class in the 

Power Structure of Great Britain," in Transactions of the Third World 
Congress of Sociology, Vol. III. London, 1 956. 

133. Emil Lederer and Jakob Marschak, "Der neue Mittelstand," in Grundriss 

der Sozialökonomik, Section IX, Part I. Tübingen, 1926. 

134. Roy Lewis and Angus Maude, The English Middle Classes. London, 1949. 

135. David Lockwood, The Blackcoated Worker. London, 1958. 

136. R. K. Merton, A. P. Gray, B. Hockey, and H. C. Selvin, eds., Reader in 

Bureaucracy. Glencoe, 1952. 

137. Q,.VJ.M'<Sk,White Collar: The American Middle Classes. New York, 195 I. 


138. Reinhard Bendix, Work and Authority in Industry: Ideologies of Manage- 

ment in the Course of Industrialization. New York and London, 1 956. 

139. Goetz Briefs, The Proletariat. New York, 1938. 

140. James Burnham, The Managerial Revolution. New York, 1 941. 

141. Theodore Caplow, The Sociology of Work. Minneapolis, 1954. 

142. Ralf Dahrendorf, "Industrielle Fertigkeiten und soziale Schichtung," Kölner 

Zeitschrift für Soziologie und Sozialfsychologie, Vol. VIII, No. 4 (1956). 

Bibliography 325 

143. Ralf Dahrendorf, Unskilled Labour in British Industry. Ph.D. Thesis. Lon- 

don, 1956. 

144. P. F. Drucker. The New Society: The Anatomy of the Industrial Order. 

New York, 1950. 

145. Henri Fayol, Administration industrielle et generale. Paris, 1 91 6. 

146. Georges Friedmann, Zukunft der Arbeit. Cologne, 1953. 

147. Peter Gay, Das Dilemma des demokratischen Sozialismus. Nürnberg, 1954. 

(Translated from The Dilemma of Democratic Socialism. New York, 

148. H. H. Gerth and C. W. Mills, "A Marx for the Managers," in 136 above. 

149. Institut für Sozialforschung, Frankfurt a.M., Betriebsklima. Frankfurt a.M., 

1954. Mimeographed. (Abridged version: Frankfurt a.M. 1956.) 

150. Heinz B. Kluth, "Arbeiterjugend — Begriff und Wirklichkeit," in Helmut 

Schelsky, ed., Arbeiterjugend — gestern und heute. Heidelberg, 1955. 

151. "Empirische Studien im Industriebetrieb," in H.-D. Ortlieb, ed., 

Hamburger Jahrbuch für Wirtschafts- und G esellschaftsfolitik, Vol. I. 
Tübingen, 1956. 

152. Jürgen Kuczynski, Die Geschichte der Arbeiter in Deutschland von lySg bis 

in die Gegenwart. Sixth ed. Berlin, 1954. 

153. Emil Lederer and Jakob Marschak, "Die Klassen auf dem Arbeitsmarkt und 

ihre Organisationen," in Grundriss der SozialökonomiJk, Section IX, 
Part II. Tübingen, 1926. 

154. "EhonMuyo, The Social Problems of an Industrial Civilization. London, 1949. 

155. W. H. McPherson, "Betrachtungen zur deutschen Arbeitsverfassung," in 

H.-D. Ortlieb and Helmut Schelsky, eds., Wege zum sozialen Frieden. 
Stuttgart and Düsseldorf, 1954. 

156. Ministry of Labour and National Service, Industrial Relations Handbook. 

London, 1953. 

157. W. E. Moore, Industrial Relations and the Social Order. New York, 1947. 

158. F. H. Mueller, Soziale Theorie des Betriebes. Berlin, 1952. 

159. National Resources Committee, "The Structure of Controls," in 35 above. 

160. Otto Neuloh, Die deutsche Betriebsverfassung und ihre Sozialformen bis zur 

Mitbestimmung. Tübingen, 1956. 

161. Andre Philip, La democratie industrielle. Paris, 195 5. 

162. Theo Pirker, Siegfried Braun, and others, Arbeiter, Management, Mitbestim- 

mung. Stuttgart and Düsseldorf, 1955. 

163. Helmut Schelsky, "Die Jugend der industriellen Gesellschaft und die Ar- 

beitslosigkeit," in Arbeitslosigkeit und Berufsnot der Jugend, Vol. II. 
Cologne, 1952. 

164. "Berechtigung und Anmassung in der Managerherrschaft," in H.-D. 

Ortlieb, ed., Wirtschaftsordnung und Wirtschaftsfolitik ohne Dogma. 
Hamburg, 1954. 

165. "Industrie- und Betriebssoziologie," in Arnold Gehlen and Helmut 

Schelsky, eds., Soziologie. Düsseldorf and Cologne, 195 5. 

166. W. H. Scott, Industrial Democracy: A Revaluation. Liverpool, 1955. 

167. F. W. Taylor, "The Principles of Scientific Management," in Scientific Man- 

agement. New York and London, 1947. 

168. ^e&tnct zndiS\<intY Wthh, Industrial Democracy. London, 191 3. 

326 Bibliography 

169. W. H. Whyte, The Organization Man. New York, 1957. 

170. "Zahlen zur Mitbestimmung," Das Mitbestimmungsgesfräch, Vol. II, No. 1 1 


171. "Die Möglichkeiten der betrieblichen Mitbestimmung nach dem Mitbestim- 

mungs- und Betriebsverfassungsgesetz," Das Mitbestimmungsgesfräch, 
Vol. II, No. II (1956). 


172. Charles Babbage, On the Economy of Machinery and Manufactures. London, 


173. A. A. Berle and G. C, Means, The Modern Corf oration and Private Proferty. 

New York, 1932. 

174. Colin Clark, The Conditions of Economic Progress. London, 1951. 

175. Barbara and J. L. Hammond, The Bleak Age. London, 1946. 

176. Eduard Heimann, Wirtschaftssysteme und Gesellschaftssysteme. Tübingen, 


177. "Kapitalismus," in Handwörterbuch der Sozialzvissenschaften, Vol. I. 

Stuttgart, Tübingen, and Göttingen, 1956. 

178. Carl Jantke, "VorindustrieUe Gesellschaft und Staat," in Arnold Gehlen and 

Helmut Schelsky, eds., Soziologie. Düsseldorf and Cologne, 1955. 

179. "IndustriegeseUschaft und Tradition," in Verhandlungen des 13. 

deutschen Soziologentages. Cologne, 1957. 

180. Hargreaves Parkinson, Ownership of Industry. London, 195 I. 

181. P. J. Proudhon, Systeme des contradictions economiques, ou Philosofie de la 

misere. 3d ed. Paris, 1867. 

182. Lionel Robbins, "The Economic Basis of Class Conflict," in 56 above. 

183. F. Rosenstiel, "Der amerikanische 'Volkskapitalismus,'" Frankfurter Allge- 

meine Zeitung, November ll, 1956. 

184. Karl Schwantag "Aktiengesellschaft: (II) AG in wirtschaftlicher Hinsicht," 

in H andwörterbuch der Sozialwissenschaften, Vol. I. Stuttgart, Tübin- 
gen, and Göttingen, 1956. 

185. Adam Smith, An Inquiry into the 'Nature and Causes of the Wealth of Nations. 

New ed. New York, 1937. 

186. Fritz Sternberg, Kapitalismus und Sozialismus vor dem Weltgericht. Ham- 

burg, 195 1. 

187. R. H. Tawney, Religion and the Rise of Capitalism. London, 1926. 

188. Andrew Ure, The Philo sofhy of Manufacturers. London, 1835. 

189. Max Weber, The Protestant Ethic and the Spirit of Capitalism. London, 1930. 


190. Hannah Arendt, The Origins of Totalitarianism. New York, 195 i. 

191. Reinhard Bendix and S. M. Lipset, "Political Sociology," Current Sociology, 

Vol. VI, No. 2 (1957). 

192. Carl Brinkmann, "Revolution," in H andwörterbuch der Sozialwissenschaften, 

Vol. IX. Stuttgart, Tübingen, and Göttingen, 1957. 

193. Anthony Downs, An Economic Theory of Democracy. New York, 1957. 

194. Th.todor'EicYitnhxi.Tg, Herrschaft der Verbände? Stuttgart, 1955. 

Bibliography 327 

195. Oscar Glantz, "Class Consciousness and Political Solidarity," in American 

Sociological Review, Vol. XXIII, No. 4 (1958). 

196. F. A. von der Heydte and Karl Sacherl, Soziologie der deutschen F arteten. 

Munich, 1955. 

197. Harold Laski, A Grammar of Politics. 6th impr. London and New Haven, 


198. Harold Lasswell, Politics — Who Gets What, When, and How? New York, 


199. S. M. Lipset, Agrarian Socialism. New York, 1950. 

200. Robert Michels, Z,ur Soziologie des Parteiwesens in der modernen Demo- 

kratie. 2d ed. Stuttgart, 1925. 

201. Talcott Parsons, "The Distribution of Power in American Society," World 

Politics, Vol. X, No. I (October 1957). 

202. Andre Philip, Le Socialisme trahi. Paris, 1957. 

203. Bertrand Russell, Power: A New Social Analysis. London, 1938. 

204. Werner Sombart, Warum gibt es in den Vereinigten Staaten keinen Sozialis' 

mus? Tübingen, 1906. 

205. J. L. Talmon, The Origins of Totalitarian Democracy. London, 1952. 


206. Ralf Dahrendorf, "Struktur und Funktion: Talcott Parsons und die Entwick- 

lung der soziologischen Theorie," Kölner Zeitschrift für Soziologie und 
Sozialfsychologie, Vol. XII, No. 4 (1955). 

207. "Out of Utopia: Towards a Re-Orientation of Sociological Analysis," 

American Journal of Sociology, Vol. LXIV, No. 2 (September 1958). 

208. Kingsley Davis, Human Society. New York, 1949. 

209. M. J. Levy, "Some Sources of the Vulnerability of the Structures of Rela- 

tively Non-Industrialized Societies to Those of Highly Industrialized 
Societies," in B. F. Hostelitz, ed., The Progress of Underdevelofed Areas. 
Chicago, 1952. 

210. David Lockwood, "Some Remarks on 'The Social System,' " British Journal of 

Sociology, Vol. VII, No. 2 (1956). 

211. R. M. Mad ver, So«<?/y. New York, 1937. 

212. Bronislaw Malinowski, A Scientific Theory of Culture and Other Essays. 

Chapel HiU, 1944. 

213- The Dynamics of Culture Change: An Inquiry into Race Relations 

in Africa. Ed. by P. M. Kaberry. New Haven, 1945. 

214. R. K, Merton, Social Theory and Social Structure. 2d ed. Glencoe, 1957. 

215. R. K. Merton and P. F. Lazarsfeld, Continuities in Social Research. Glencoe, 


216. Talcott Parsons, The Structure of Social Action. 2d ed. Glencoe, 1949. 

217. "The Present Position and Prospects of Systematic Theory in Soci- 
ology," in Essays in Sociological Theory. Rev. ed. Glencoe, 1 954. 

218. K. R. Popper, Logik der Forschung. Vienna, 1935. 

219. The Ofen Society and its Enemies. 2d ed. (rev.). London, 1952. 

220. A. R. Radcliffe-Brown, "On the Concept of Function in Social Science," in 

Structure and Function in Primitive Society. London, 1952. 

328 Bibliography 

221. A. R. RadcliflFe-Brown, "On Social Structure," in Structure and Function in 

Primitive Society. London, 1952. 

222. Bertrand Russell, "Structure," in Human Knowledge, Its Scofe and Limits. 

London, 1948. 


223. Henri Bergson, "Die Wahrnehmung der Veränderung," in Denken und 

schöfjerisches Werden. Meisenheim, 1948. 

224. Ralf Dahrendorf, Marx in Persfektive: Die Idee des Gerechten im Denken 

von Karl Marx. Hannover, 1953. 

225. "Philosophie und Soziologie bei Karl Marx: Marx-Kritik und ihre 

Bedeutung," Geist und Tat, Vol. X, No. 3 (1955). 

226. G.W.Y.YiQgd, Phänomenologie des Geistes. 5th ed. Leipzig, 1949. 

227. Max Horkheimer and T. W. Adorno, Dialektik der Aufklärung. Amsterdam, 


228. KsltI Mannheim, Ideology and Utofia. London, 1946. 

229. Elisabeth Noelle and E. P. Neumann, Jahrbuch der öf entlichen Meinung, 

1947 bis I9S5- Allensbach, 1956. 

230. David Riesman, The Lonely Crowd. New Haven, 1 950. 

231. Helmut Schelsky, Wandlungen der deutschen Familie in der Gegenwart. 

3d ed. Stuttgart, 1955. 

232. Alexis de Tocquerille, Democracy in America. Ed by P. Bradley. Ncvr York, 


Subject Index 

Arbitration, 65f, ^^']n■, zzgiy 256, 26of, 
277, 308 

Associations, imperatively coordinated, 
138, i67f, I72f, 197, 213, 237; in- 
dustrial, I2S, 2lff, 41-48, 48-51, 

78f, no, 142, 235f, 248-56, 296; 
political, 9off, 95, 197, 235, 289-300 
Authority, sge Power 

Bourgeoisie, seg Ruling class 

Bureaucracy, 6, 255, 295-301 ; charac- 
teristics of, 56, 296; place of, in class 
structure, 55, 56f, 9off, 99, 256, 258, 
296, 299f; power and, 56, 9 if, 94, 99, 
I07f, 256, 295, 297, 298f; role of, 
in class conflict, 300f, 308f ; white col- 
lar and, distinction of, 55, 255, 295 

Canada, 3 1 On 

Capital (capitalists, capitalist class) : com- 
position of, I of, 15, I9f, 32f, 80, 242; 
decomposition of, 2 if, 41-48, 79, 84, 
87 ff, 93 f, 254; emergence of, 4ff ; see 
also Ruling class 

Capitalist society: class conflict and, 65, 
112, 113, 125, 242ff ; concept of, 36- 
41, 4iff'; Marxian view of, lofi", 30, 
32-34, 131; place in history, 28; 
structure of, 2iff, 68f, 80, 85, no 

Caste, 95, 298 

Change, social: analysis of, 26f, ii3ff, 
119-24, I32f, 159, 160, i6iflF, 231; 
class conflict and, 19, 20, 2 5f, 28, 35, 
76, i24flF, 126-30, 152, 154, 207, 
208, 231-36, 237, 240, 245f; con- 
cept of, 32, 127, I3if, 238; opera- 
tional definition of, 2320"; radicalness 
of, 35, 130, 234, 235, 236, 238, 240, 

278, 314-18; suddenness of, 130, 
234f, 236, 238, 240, 278, 307-14; 
types of, 124, i27flF 

Citizenship: class and, 63 f, loif; devel- 
opment of, 6 iff, I oof; industrial, 65, 
III; political, 292f 

Class: power (authority) and, 6f, 83f, 
107, I36f, I48f, 165, 172, I93f, 238, 
285f; property and, 6f, I if, 21, 22f, 
33, 83f, 87f, 89f, 93, 137, 141, 172, 
242, 254; psychology of, 14, 74f, 76, 
114, 144-48, I74f, lySf, 189-93, 
196, 280-89; stratification and, 4, 7, 
19, 63f, 74f, 76f, 86, 91, 97, 99n, 
107, 129, i39f, 171, 215-18, 253f, 

Class, concept of: definition of, 6f, I36f, 
I38ff, i5off, I72f, 179, 189, 201-5, 
238, 245, 247; dilution of, 74-77, 86, 
91, 99n, I45ff, I50ff; heuristic pur- 
pose of, 9f, 19, 126, 15 if, 203, 237; 
history of, 3f, 7, 202f; Marxian, 10- 
18, 20-25, 30ff, 77f, i33f, i36ff, 245 

Class conflict, theory of, i6ff, 2 5f, 84f, 
9Sf, 130-33, 133-36, I53f, 210-36, 

Class consciousness, 25, 146, 148, 179, 
190, igiff, 238f, 281-89 

Class formation, theory of, 11- 16, 23ff, 
153, 172-89, 238f 

Class interests: latent, 178, 237; mani- 
fest, I78f, I90f, 238; Marxian notion 
of, I4ff, 23 f; "objective" nature of, 
15, 23 f, I74f, l76ff; substance of, 15, 
126, i75f, 179, 242 

Class theory: elements of, 8f, 18-27, 27— 
33, 150-54, l64f, 237; limitation« of, 
3^, 41, 97^» lOS. 126-30, 152, 204; 


Subject Index 

task of, 19, 73 f, 86, iijf, 119, i24flF, 

151, 152, 237 

Classless society, 18, 28, 29, 86, 2i8ff 

Co-determination in German industry, 
66, 261, 263-67, 279, 285 

Coercion model of society, 27, 96, 114, 
158, 159, i62ff, 170, 237, 284, 287f 

Communist society: modern sociological 
theory and, 27n; place of, in Marx's 
philosophy of history, 28; structure of, 
22, 30n, 82, 83; see also Classless so- 
ciety, Totalitarian society 

Conciliation, ösf, 2 2 7n, 228, 230, 256, 
259, 260, 308 

Conflict, social: concept of, I35f, 2085", 
238; functions of, I25f, 206— lO; in- 
tensity of, 3 if, I33ff, 210-13, 215, 
218, 222f, 239, 242, 243, 245, 277f, 
279, 314-18; regulation of, 64-67, 
134, 154, 223-31, 239, 257-67, 
307—14; types of, 129, 214; ubiquity 
of, 162, 208, 219, 224n, 237; violence 
of, I33ff, 210-13, 2i4f, 2i7f, 223, 
230, 239, 244, 245, 277, 279, 307- 
14; see also Class conflict 

Conflict group, see Class 

Democracy, 228; industrial, 64—67, 95, 
99, 244, 257-67; political, 95, 234, 
244, 260, 293, 307-14, 318 

Denmark, 98 

Dichotomy of society: in terms of power, 
71, 169, i7of, 173, 251, 292f, 296; 
in terms of two-class model, I9f, 33f, 
52f, 97, 126, 199; as viewed by peo- 
ple, 282-89, 297f 

Education and social mobility, 59, 105; 
in capitalist society, 33f, 58f, 243f, 
245; in industrial society, 57flf, 80; in 
post-capitalist society, 59, 105, 315; in 
totalitarian society, 3i5f; see also Mo- 
bility, social 

Elites, see Ruling class 

England, 37, 41, 50, 53, 58, 59, 80, 105, 
141, 187, 217, 225, 230, 243, 26on, 
261, 262, 263, 265, 275, 278, 281, 
285, 287 

Equality: class and, 63f, lOif, 105, 107; 

of rights, 6 if, 69, lOOf, I05f, 23 1; 

of status, 63, 99, 104, I05f, 219 
Estate, 5n, 6f, 7n, 17, 18, 104, 288 
Europe, 39, 67, 1 37, 199, 230, 235, 270, 


Feudal society: place in history, 28; struc- 
ture of, 4f, 5f 

France, 41, 53, 58, 183, 199, 262, 3o6n, 

French Revolution, 17, 135, 269 

Germany, 41, 42, 50, 53, 58, 98, 105, 
106, 132, 142 161, 162, 217, 219, 
23011, 233n, 243, 259, 261, 262, 263, 
264, 265, 266, 267, 281, 283n, 285, 
286n, 309n, 3 ion, 3i3n 

Group conflict, see Class conflict. Conflict 

Holland, 129 
Hungary, 132, 224 

Images of society: analytical device, 26, 
Ii2f, 115, 157-65, 206f; empirical 
finding, 280-89, 298 

Industrial conflict, 64f, 99, 228fi", 241- 

Industrial revolution, 4f, 47, 68, 135, 

Industrial society: applicability of theory 
of class to, 152, l86n, 200, 246f; class 
conflict in, 64, 103, 104, 107, 131, 
193, I95f, 197, I99f; compared with 
feudal society, 5ff^; concept of, 36—41, 
4ifi", 1 1 of; structure of, 570", 6 if, 
68f, 70, 80, 90, I42f, i87f 
Industry, see Associations, industrial 
Industry and society: relations of, I2ff", 
23, 88ff, l40-44> 243> 3i6f; separa- 
tion of, I43f, i7if, 214, 245, 267-76, 


Integration model of society, 96, Il3f, 
158, 159, i6off, i63f, 170, 210, 284 

Intellectuals, 8of, 186 

Interest groups: class and, I79f, 202, 
204f, 238, 246f; concept of, l8of, 
238; quasi-groups and, 24f, 1 8 if, 
182-89, 190, 226, 238f, 242f 

Interests, see Class interests 

Israel, 218, 2 1 9, 221 

Subject Index 


Italy, 58 
Japan, 58 

Labor (working class) : composition of, 
lof, 32f, 51, 242; decomposition of, 
48—5 I ; emergence of, 5f 

Managers (managerial class) : emergence 
of, 2 if, 42f, 84, l6lf; place in class 
conflict, 48, 88, 99, 254, 264f; social 
characteristics of, 44f, 46f, 87flF 

Mediation, 65f, 227n, 228f, 230, 256, 
26of, 277, 308 

Methodological problems relating to: 
analysis in terms of structure (institu- 
tions) and value (norms), 7on, 96, 
231 ; assumptions, 23, 46n, I I2f, 173 ; 
definitions, 30, 38n, 39, 137, 138; 
description, 38n, 39, 124; heuristic 
purpose, 19, 75, 124, 125, 203; so- 
ciology and philosophy, i, 8, 27ff", 
118; theory, 35, 72, i5of, 152, I53f, 
l83n, 241 

Middle class, 6, 50-57, gif, 94, 103; 
components of, 52f, 55f; development 
of J 33f> 5^5 68, 98; image of society 
and, 283f, 286f, 298; Marxist view of, 
79ff ; role of, in class conflict, 53f, 95, 
98f ; values and styles of life, 63, 104; 
voting behavior of, 53; see also Bu- 
reaucracy, White-collar woricers 

Mobility, social: class stability and, 17, 
57, 60, 103, 105, I08f, 144, 148, 
191, 218-23; classless society and, 
108, 218—23; intensity/violence of 
conflict and, 60, 2 2 2f, 239, 3 i 5f ; rad- 
icalness/ suddenness of change and, 
6of; ruling groups and, 46, 80, 92, 
232flF, 265f, 296, 310; sociological 
analysis of, 57flf, 58n, 69; see also Edu- 
cation and social mobility 

Monism (superimposition) and intensity 
/violence of conflict, 140, 193, 212— 
15, 215-18, 239, 243f, 245, 3i6flF 

Pluralism (dissociation) : freedom and, 
3l7f; intensity /violence of conflict 
and, 140, 193, 212-15, 215-18, 239, 
245, 267f, 277f, 3i6flF 

Poland, 132 

Political conflict, 66, 99, 2 2 8ff, 243 f, 

Political parties: classes as, 17, 25 ; demo- 
cratic, 304f, 306, 307, 309f, 31 in; 
democratic, as interest groups, 1 8 1 f, 
188, 246; democratic, and personality 
types, 113; right to form, 62; totali- 
tarian, 304, 305, 311, 313, 3i6f; 
trade unions and, 271, 274, 275; vot- 
ing of middle class and, 53, 56, 301 

Post-capitalist society: images of society 
in, 28 iff; industrial conflict in, 65f, 
99, 252f, 254flF, 257fl^, 27iff", 277, 
279, 318; Marxian class model and, 
97, 102, 105 ; pluralism vs. monism in, 
268ff, 275, 277f, 318; political con- 
flict in, 4ifl^, 48ff, 71, 94, 95f, I07f, 
253f, 29ofF, 292flF, 297flt, 303, 30401, 
3iof; structure of, 5 iff, 59, 87f, 98f, 
i02ff, 105, I36f, 254, 274, 315 

Power: concept of, 165-73, 237; concept 
of, as determinant of class, 83f, 107, 
I36f, I48f, 165, 172, I93f, 238, 
285f; distribution of, 107, I70f, I95f, 
249, 29of, 292, 296; division of labor 
of, 53f, 56, 9 if, 94, 246, 25 5f, 297ff; 
industrial (economic), i if, I37ff, 141, 
I43f, 214, 248-57, 271, 273f, 275f; 
political, I2ff, I37ff, 141, I43f, 214, 
271, 273f, 275f, 289-95; political, in 
preindustrial society, 5f ; property and, 
I If, 21, 22f, 33, 83f, 87f, 89f, 93, 
137, 141, 172, 242, 254; social strati- 
fication and, 215-18, 242, 316, 317; 
as substance of class interests, 176, 232, 
252; universality of, 71, i68f, 2l8f, 
248f, 251, 29if 

Proletariat, see Subjected class 

Property: alienation and, 2 8f, 30 ; in cap- 
italist society, 37, 38, 39, 40, 85f, 
242; in communist society, 82f, 304, 
316; in post-capitalist society', 96, 99, 
292f, 304f, 3o6f, 317; in preindus- 
trial society, 5f; as determinant of 
class, I if, 20-23, 31, 33, 77f, 83f, 
85f, 136— 41 ; ownership and control, 
separation of, 41-48, 68, 78f, 84, 87f, 
93f, 99, 254; power and, iif, 21, 22f, 


Subject Index 

33, 83f, Syf, ggf, 93, 137, 141, 172, 
242, 254 

Quasl-groups: class and, I79f, 202, 204f, 
238, 246f; concept of, 180, 237f; in- 
terest groups and, 24f, 18 if, 182-89, 
190, 226, 238f, 242f 

Revolution: Burnham's, Croner's, Druck- 
er's concept of, 87ff, 9 if, logf; class 
conflict and, 9, 15, 18, 26, 35, 51, 
i30-33> i99n> 221, 224, 230, 244; 
concept of, 83, 130, 233, 234n, 299, 
308f, 31 1; Marx's theory of, 32, 125, 


Role analysis: in structural - functional 
theory, 120; of bureaucratic positions, 
296f, 30of; of class membership, 
l48flF, 190, I92f; of class structure, 
145, I47f; of interests, 148, 175, 
I77f, 237; of managerial positions, 47, 
265, 266; of power, 1 7 if, 249f 

Rules of the game (in conflict regulation), 
226f, 229f, 260, 308, 309, 3 ion 

Ruling class, elite theory of, 194, 195, 
196, I98f, 222, 3iof 

Ruling class: general features of, 14, 17, 
99, I70f, 174, 176, 193-98, 218; 
in capitalist society (bourgeoisie), I2f, 
15, 34f, 24.2S, 247; in post-capitalist 
society, 89f, 251, 254, 256, 271, 27sf, 
292flF, 297, 298, 299, 300, 301-7; 
in totalitarian society, 82ff, 304, 305, 
3i2f, 3i6f 

Socialist society, 57, 8of, 85, 86 
Stratification (stratum), social: class and, 
confusion of, 4, 74f, 76, 86, 91, 99n, 
107, 129, 139, 253f; class and, dis- 
tinction of, 7, 19, 63 f, 76f, 97, 107, 
I39f, 171, 215-18, 283; intensity/ 
violence of conflict and, 140, 239, 242, 
245, 3i6ff 

Structural-functional analysis: assumptions 
and tenets of, 27, 112, iigi, 158, 159, 
161, 207; critique of, 114, 120-23, 
131, I59f, i6lf, 210 

Status, social: concept of, 6f, 1 00, 217; 
inequality of, 7of, lOl, 106, I39f, 
217; leveling of, 63 f, loof, 103, 104, 
I05f, 219 

Subjected class: general features of, I70f, 
176, 198-201, 218; in capitalist so- 
ciety (proletariat), I2f, 14, 15, 34f, 
13 if, 242ff, 247; in post-capitalist so- 
ciety, 94f, 251, 253f, 255f, 271, 275f, 
292f, 308, 309f ; in totalitarian society, 
3i2f, 3i6f 

Switzerland, 281, 283n, 285, 286n 

Totalitarianism, 57, 83, 221, 225n, 279, 

Totalitarian society: mobility in, 313, 
3i5f; pluralism vs. monism in, 304f, 
3l6f; regulation (suppression) of con- 
flict in, 186, 224, 3i2f, 314, 315 

Trade unions: development of, 62, 243, 
279; in industrial conflict, 225, 244, 
260, 263, 264n, 277; as interest 
groups, i8if, 246, 252f, 258f; leaders 
of, 112; middle class and, 56; political 
parties and, 271, 274, 275 

United States, 17, 37, 39, 41, 42, 44n, 
50, 57, 58, 60, 67, 105, no, 128, 
137, 142, 143, i69n, 215, 217, 225, 
230, 233n, 259, 261, 262, 263, 265, 
270, 275, 278, 281, 286, 289, 291, 

USSR, 77, 78,- 80, 81, 82, 188, 222n, 
225n, 233, 315, 316 

Veto groups, 306, 307 

White-coUar workers, 6; bureaucracy and, 
distinction of, 55, 255, 295; place of, 
in class structure, 55f, 25 5f, 295 

Working class, se^ Labor 

Author Index 

Adorno,Th.W., 63, 328 
Albrecht, Gerhard, 319 
Angell, R. C, 322 
Arendt, Hannah, 326 
Aristotle, 157 

Aron, Raymond, 193, 194, 195, 196, 

Babbage, Charles, 48, 326 

Bahrdt, H.P., 37, 55,281,321,324 

Barnard, C. J., 322 

Bauer-Mendelberg, Käthe, 319 

Ben-David, Joseph, 218, 221, 324 

Bendix, Reinhard, 3, 8, 44, 45, 52, 53, 
74, 76, 89, 90, 117, 119, 139, 166, 
203, 221, 250, 254, 261, 296, 298, 
299» 301, 3"> 320, 321, 323, 324, 

Bergson, Henri, 122, 131,328 

Berle, A. A., 42, 326 

Bernard, Jessie, 322 

Bernstein, Eduard, 84 

Bolte, K. M., 5 8, 2 1 7, 3 23 

Bonham,John, 53, 324 

Bottomore, T. B., 37, 63, 3 1 5, 320 

Bowlby, John, 208, 322 

Braun, Siegfried, 325 

Briefs, Götz, 324 

Brinkmann, Carl, 130, 320, 326 

Burnham, James, 43, 87-91, 92, 93, 94, 
99, 109, 115, 136, 137, 141, 144, 
192, 254, 268, 271, 272, 276, 302, 

Caplow, Theodore, 50, 324 
Carr-Saunders, A. M., 49, 322 
Centers, Richard, 75, 146, 147, 189, 
281, 283, 286, 287, 288, 289, 320 

Clark, Colin, 79, 142, 326 

Cole, G. D. H., 117, 222, 320 

Coleman, J. S., 322 

Corey, Lewis, 324 

Coser, L. A., 207, 208, 209, 226, 313, 

Cox,0. C, 6, 7, 74, 320 
Croner, Fritz, 52, 53, 74, 75, 91-92, 93, 

94, 109, 146, 147, 256, 296, 324 
Crozier, Michel, 53, 324 

Davis, Kingsley, 59, 63, 322 

Djilas, Milovan, 77, 78, 81, 82-84, 86, 

93, 109, 115, 136, 137, 141, 304, 

Downs, Anthony, 309, 326 
Drucker, P. F., 49, 109-14, 115, 142, 

158, 252, 262, 325 
Dubin, R., 206, 208, 226 
Durkheim, Emile, 67, 1 19 

Engels, Friedrich, 4, 8, 30, 3 19 
Eschenburg, Theodor, 326 

Fahlbeck, Pontus, 3 1 9 
Fayol, Henri, 49, 325 
Fedoseyev, P. N., 78, 79, 81, 320 
Ferguson, Adam, 4, 5, 29, 319 
Floud, Jean, 323 
Fourier, Charles, 4 
Freud, Sigmund, 177, 178 
Friedmann, Georges, 49, 325 

Gay, Peter, 325 

Gehlen, Arnold, 325, 326 

Geiger, Theodor, 8, 19, 20, 36, 43, 49, 
52, 54, 64, 65, 66, 74, 84, 90, 97- 
100, 102, 109, 115, 117, 119, 137, 


Author Index 

139, 141, 146, 147, 150, 151, 174, 
175, 176, 178, 179, 195, 203, 275, 

Gellius, 4 

Gerth, H. H., 90, 325 

Ginsberg, Morris, 175, 180, 182, 192, 

202,203, 321 
Glantz, Oscar, 289, 327 
Glass, D. v., 58, 80, 105, 217, 323 
Gordon, M. M., 321 
Gouldner, Alvin, 296, 324 
Grey,A. P., 324 
Gross, Llewellyn, 321 
Gurvitch, Georges, 1 30, 203, 32 1 

Halbwachs, Maurice, 32 1 

Hall,J.R., 323,324 

Hammond, Barbara, 6, 326 

Hammond, J. L., 6, 3 26 

Hegel, G. W. F., 8, 20, 27, 29, 62, 85, 

125, 131, 132, I57>328 
Heimann, Eduard, 38, 326 
Henderson, A. M,, 320 
Heraclitus, 125 

Heydte, F. A. von der, 53, 327 
Hilferding, Rudolf, 44 
Himmelweit, Hilde T., 323 
Hobbes, Thomas, 157 
Hockey, B., 324 
Hoggart, Richard, 189, 281, 282, 283, 

Horkheimer, Max, 63, 328 
Hostelitz, B. F., 327 
Hume, David, 29 

Inkles, Alex, 323 

Janowitz, Morris, 60, 220, 323 
Jantke, Carl, 326 
Jones, D.C., 49, 322, 323 
Jiires, E. A., 281, 321 

Kaberry, P. M., 327 

Kant, Immanuel, 157 

Kapadia, K. M., 322 

Kautsky, Benedict, 77 

Kelsall, R. K., 324 

Kerr, Clark, 66, 226, 227, 229, 322 

Kesting, Hanno, 281,321 
Kluth, Heinz, 50, 325 
Kolb,W.L., 322 
Kornhauser, A. W,, 322 
Kuczynski, Jürgen, 84, 325 

Landshut, Siegfried, 218, 321 

Laski, Harold, 197, 290, 291, 293, 302, 

Lassalle, Ferdinand, 130 
Lasswell, Harold, 166, 327 
Lazarsfeld, P. F., 327 
Lederer, Emil, 51,52,324,325 
Lenin, W. I., 78 
Lenski, G. E., 321 
Levy, M, J., 120, 127, 327 
Lewis, Roy, 52, 324 
Lipset, S. M., 3, 8, 5 8, 74, 76, 1 1 7, 1 1 9, 

139, 203, 310, 311, 320, 321, 323, 

Locke, John, 29 
Lockwood, David, 52, 66, 159, 160, 163,, 

229, 261, 296, 298, 322, 324, 327 
Lukacs, Georg, 77, 179, 319 
Lunt, P. S., 75, 323 

Machiavelli, Nicolo, 195 

Maclver, R. M., 75, 181, 327 

Mack, Raymond, 209, 21 1 

Mackenroth, Gerhard, 323 

McPherson, W. H., 260, 262, 325 

MacRae, D. G., 322 

Malinowski, Bronislaw, 128, 181, 184, 

185,259, 327 
Mannheim, Karl, 80, 328 
Markmann, Heinz, 322 
Marschak, Jakob, 51, 52, 324, 325 
Marshall, Alfred, 1 19 
Marshall, T. H., 51, 61, 62, 63, 65, 74, 

97, 100-109, 115, 127, 139, 144, 
146, 147, 192, 203, 209, 231, 260, 

Marx, Karl, 3-35, 36, 37, 38, 39, 41, 
42> 43> 44> 47. 48, 50> S^» 54» 56, 
57> 58, 59> 60, 61, 62, 65, 66, 69, 72, 
73. 74 75. 76, 77. 78, 81, 83, 84, 85, 
86, 87, 88, 89, 90, 91, 92, 93, 95, 97, 

98, 99, 102, 103, 107, 108, 109, 1 1 1, 

Author Index 


113, 115, 116, 117-54, 158, 160, 

172, 174, 175, 176, 179, 183, 184, 

185, 186, 187, 188, 192, 194, 195, 

198, 199, 200, 201, 202, 203, 204, 

212, 215, 218, 230, 241, 242, 243, 

244, 245, 246, 248, 251, 254, 268, 

Maude, Angus, 52, 324 

Mayer, K. B., 321 

Mayo, Elton, 69, 109-14, 115, 158, 

Means, G. C, 42, 326 
Merton, R. K., 72, 90, 120, 178, 296, 

Michels, Robert, 90, 92, 279, 302, 311, 

Millar, John, 4, 5,29,319 
Mills, C. W., 43, 52, 54, 90, 169, 170, 

Mombert, Paul, 74, 320 
Moore, W. E., 49, 63, 66, 227, 251, 

Mosca, Gaetano, 166, 1 93, 194, 195, 

196, 197, 198, 199, 200, 203, 220, 

222,255, 302, 320 
Moser, CA., 324 
Mueller, F. H., 250, 251, 325 
Mukherj ee, Ramakrishna, 58, 324 

Nemchinov, V. S., 77-82, 84, 321 
Neuloh, Otto, 250, 267, 325 
Neumann, E. P., 328 
Newman, Frank, 208 
Noelle, Elisabeth, 328 

Ortlieb, H.-D., 320, 321, 325 
Ossowski, Stanislaw, 82, 287, 288, 289, 

Pareto, Vilfredo, 119, 166, 193, 194, 
195, 196, 197, 198, 199, 200, 203, 
Parkinson, Hargreaves, 42, 326 
Parsons, Talcott, 112, 119, 120, 122, 
123, 126, 131, 134, 140, 144, 159, 
160, 161, 163, l66, 167, 169, 170, 
175, 177, 210, 2i8, 231, 267, 320, 

Pear,T. H., 322 

Pfautz,H.W., 74, 75,323 

Philip, Andre, 50, 262, 325, 327 

Pirker, Theo, 265, 325 

Plato, 157 

Popitz, Heinrich, 189, 281, 282, 283, 

284, 285, 286, 287, 288, 289, 290, 

Popper, K.R., 73, 100,327 
Proudhon, P. J., 48, 326 

Radcliffe-Brown, A. R., 121, 163, 327, 

Rathenau, Walther, 87 
Ratzenhofer, 175 
Renner, Karl, 19, 43, 44, 53, 84, 89, 

93-96, 109, 115, 119, 130, 136, 141, 

144, 148, 160, 168, 203, 254, 256, 

Ricardo, David, 4 
Riesman, David, 291, 292, 300, 304, 

Robbins, Lionel, 175, 326 
Roethlisberger, F. J., 262 
Rogoff, Natalie, 58, 324 
Rosenstiel, F., 42, 326 
Rossi, P. H., 323 
Rousseau, J. J., 95, 157, 164 
Russell, Bertrand, 120, 121, 166, 328 

Sacherl, Karl, 326 

Saint-Simon, C. H. de, 4 

Schelsky, Helmut, 59, 62, 63, lOO, I02- 
9, 115, 246, 250, 261, 262, 272, 273, 
290, 291, 292, 321, 323, 324, 325, 

Schumpeter, J. A., 31, 40, 84-87, 93, 
109, 117, 119, 136, 139, 141, 144, 
148, 149, 150, 151, 153, 191, 192, 

Schwantag, Karl, 42, 326 

Scott, W.H., 325 

Selvin, H. C, 324 

Sering, Paul, 40, 43, 46, 86, 87, 322 

Sheppard, H. L., 1 12, 1 14, 322 

Simmel, Georg, 1 92, 207, 211, 226 

Small, Albion, 175 

Smith, Adam, 4, 48, 326 


Snyder, Charle 
Sombart, Wer 

Sorel, Georges 
Sorokin, Pitiri) 
Spann, Othma 
Stalin, J. W., 5 
Sternberg, Fri 
Sumner, W. G 

Talmon, J. L. 
Tawney, R. f: 

Taylor,F.W.,49, 251,252, 325 
Thomas, G., 59,278,324 
Tocqueville, Alexis de,6i, 231,32? 
Torgersen, Ulf, 1 1 3 
Tropp, Asher, 296,324 


t, 40, 49> 
143, i66, 
237, 290, 

Willener, Alfred, 189, 281, 282, 283, 

Wilson, Logan, 322 

Zetterberg,H.L., 58, 323 
Ziegel, W., 323 

Graduate Reserve Service 
University Research Library 
liOs Angeles. Calirornia 90024 

]4r •■ 

1216 8455 

1^ So ^oo 


Snyder, Charles, 209, 211 

Sombart, Werner, 6, 37, 38, 57> 3^0, 

Sorel, Georges, 207 
Sorokin, Pitirim, 74, 320 
Spann, Othmar, 320 
Stalin, J. W., 81, 134 
Sternberg, Fritz, 270, 326 
Sumner, W.G., 175 

Author Index 

Ure, Andrew, 4, 48, 326 
Veblen, Thorstein, 87, 302, 319 

Talmon, J. L. 
Tawney, R. ¥. 
Taylor, F. W. 
Thomas, G., t 
TocqueviUe, I 
Torgersen, U 
Tropp, Asher. 

Warner, W. L., 75, 76, 189, 322, 323 
Webb, Beatrice, 325 
Webb, Sidney, 325 
Weber, Alfred, 80, 322 
Weber, Max, 6, 7, 24, 37, 38, 40, 49» 
68. 6q. QO. q2, liq, 136, 143, 166, 

237, 290, 

University of California Library 
Los Angeles 

This book is DUE on the last date stamped below. 





0, 326 


282, 283,